The International Simutrans Forum

Community => Game Servers => Topic started by: jamespetts on January 14, 2012, 09:03:19 PM

Title: [Closed] New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on January 14, 2012, 09:03:19 PM
Please note that this thread relates to the original game played on this server, now closed. A new game on this server will be posted shortly. In the meantime, the final state of the saved game can be downloaded here (http://www.bridgewater-brunel.me.uk/saves/bridgewater-brunel-game1-final.sve).




Server status

See the listing server (http://servers.experimental.simutrans.org/list) page.

Program version

10.11

Pakset version

Pak128.Britain-Ex 0.8.3

***

I have rented some rackspace and set up a Simutrans-Experimental server (using Simutrans-Experimental 10.5 and Pak128.Britain-Ex 0.8.3). It is a large map, starting in 1825: The server has fairly high specifications, including a dedicated 3Gb of RAM, so should be able to cope with the map size, especially since it will only be running one game at a time.
 
This server, at bridgewater-brunel.me.uk, is now listed on the Experimental announce server (http://servers.experimental.simutrans.org) (thanks to Timothy for producing one for Experimental).
 
 The server is set to pause when no clients are connected. Below is a screenshot of the map, at a 3:1 zoom ratio.

(http://simutrans-germany.com/files/upload/jan-2012-1-server.png)
 
This is the first public Simutrans-Experimental server, so I should be very grateful for any feedback, especially as to how this works in comparison to Simutrans-Standard in a multi-player setting. There might also be unexpected issues, and it may be necessary to restart the map at short notice as a result, until such issues as might arise are resolved.
 
I shall not presently specify any rules, as such, other than a general request for players to be respectful and polite to others. Both co-operative and competitive play is permitted. Multiple players may share a single transport company if they wish. I shall play as the public service player in this initial game, and shall, in so far as my time permits, try to act as a government might, with the aim of maximising overall population growth within the constraints of the public service player. I might also play separeately as an individual transport company if I get the time.
 
All slots are free apart from observer and public service player, so do get stuck in!
 
Title: Re: New Simutrans-Experimental server
Post by: prissi on January 14, 2012, 10:13:18 PM
I would strongly suggest a smaller map without trees. Your map is about 3 MB to download, and the server need already a lot of time to save it. Every time a player connects, it will take 1-2 min to continue to play.

And a lower fps than 25, maybe 10. My computer is an Athlon2 X64 6000+ (still very fast concerning single core perfomance) and runs at 80% on one core on this map. With any trains or routing, it will almost desync immeadiately.
Title: Re: New Simutrans-Experimental server
Post by: rsdworker on January 14, 2012, 10:19:00 PM
sounds great i will join to play so its online now?
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 14, 2012, 10:47:38 PM
Prissi,

thank you for the tips. I shall set the fps to 15 for future games. I shall leave this as it is for the time being unless instability and desyncing is reported. May I ask those who have played it so far - has it been stable or have there been frequent desyncs?

As to the map size and trees, I really did want to be able to have be able to have a large map server, as it really makes a difference to gameplay having a larger map because of the scale. There is a tradeoff between having a small, fast loading map and a larger but slower loading map. My tests showed that I can connect to this map in just under 45 seconds.

rsdworker - it is indeed online now - do feel free to join!
Title: Re: New Simutrans-Experimental server
Post by: rsdworker on January 14, 2012, 11:00:32 PM
Prissi,

thank you for the tips. I shall set the fps to 15 for future games. I shall leave this as it is for the time being unless instability and desyncing is reported. May I ask those who have played it so far - has it been stable or have there been frequent desyncs?

As to the map size and trees, I really did want to be able to have be able to have a large map server, as it really makes a difference to gameplay having a larger map because of the scale. There is a tradeoff between having a small, fast loading map and a larger but slower loading map. My tests showed that I can connect to this map in just under 45 seconds.

rsdworker - it is indeed online now - do feel free to join!
yeah i did join but i crashed because its was lost syc
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 14, 2012, 11:07:41 PM
May I ask - what were you doing when you lost synchronisation?
Title: Re: New Simutrans-Experimental server
Post by: rsdworker on January 14, 2012, 11:15:10 PM
May I ask - what were you doing when you lost synchronisation?

i was exploring and setting up company - now i have track setup now - my company name is City tramway
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 14, 2012, 11:23:58 PM
Did you desync again?
Title: Re: New Simutrans-Experimental server
Post by: rsdworker on January 14, 2012, 11:24:40 PM
Did you desync again?

nope i left because i was discrated by rl i will be back in soon
Title: Re: New Simutrans-Experimental server
Post by: Ashley on January 15, 2012, 12:30:58 AM
What's not working about the announce server?
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 15, 2012, 12:41:09 AM
It's not announcing. The list is empty, even though the server has been running for longer than the announce interval (and I have set it to announce manually using nettool), and announce=1 is set in simuconf.tab.
Title: Re: New Simutrans-Experimental server
Post by: Guvnor on January 15, 2012, 01:02:10 AM
James,

I will join if it's okay.

Dave
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 15, 2012, 01:08:31 AM
It's certainly okay - feel free! I'd be very interested in any feedback, both as to stability/performance and to how Experimental gameplay works in comparison to Standard in multiplayer.
Title: Re: New Simutrans-Experimental server
Post by: sdog on January 15, 2012, 03:07:48 AM

James, congratulations for your new server!
(i'd connect immediately, if i wasn't afraid of getting completely and utterly addicted -- again.)

I would strongly suggest a smaller map without trees. Your map is about 3 MB to download, and the server need already a lot of time to save it. Every time a player connects, it will take 1-2 min to continue to play.
[...]
(bold by me)

Instead of saving the location of trees, perhaps only the parameters to generate the forrests could be saved and the trees generated on loading the map. The trees are only eyecandy, if they are not persistent and identical for all players in a network game, it shouldn't really matter.

@james, you could also generate maps that are much longer than wide, you get the long distance interaction, without increasing the map area too much.

Player planted trees, could be excepted of course.
Title: Re: New Simutrans-Experimental server
Post by: Guvnor on January 15, 2012, 04:57:24 AM
Server down?
Title: Re: New Simutrans-Experimental server
Post by: Ashley on January 15, 2012, 10:04:33 AM
It's not announcing. The list is empty, even though the server has been running for longer than the announce interval (and I have set it to announce manually using nettool), and announce=1 is set in simuconf.tab.

I can't see any announce attempts in the log file. Can you do some further testing of announce settings? If nothing appears in the log it probably means Simutrans isn't sending the announce.
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 15, 2012, 12:14:31 PM
Guvnor - apologies for the server being down. That was my error, as I did not realise that closing the terminal session when I shut down my own computer would terminate the Simutrans session! I have fixed that now by using Timothy's scripts - the server should be back up.

Timothy - hmm, I am not sure why it is not announcing. Can I check that I have the various settings right in simversion.h? They are reproduced below:

Code: [Select]
/*********************** Settings related to network games ********************/

/* Server to announce status to */
#define ANNOUNCE_SERVER "servers.experimental.simutrans.org:80"

/* Relative URL of the announce function on server */
#define ANNOUNCE_URL "/announce"

/* Relative URL of the list function on server */
#define ANNOUNCE_LIST_URL "/list?format=csv"

/* Name of file to save server listing to temporarily while downloading list */
#define SERVER_LIST_FILE "serverlist.csv"

#endif

Are these correct?

Edit: Also, I am having trouble using a non-root user to run Simutrans - I have set up a "simd" user, but when I run Simutrans with this user, it cannot accept any incoming connexions. Do you have any idea what to do about this?

SDog - actually, the trees don't seem to be causing much difficulty so far. For a high-specification map, a 45 second load time is fairly reasonable, I think. We shall have to see with experience whether it is unserviceable in play or not, but I really do want to test the boundaries of what can be done with high specification maps on a high specification server.
Title: Re: New Simutrans-Experimental server
Post by: ojii on January 15, 2012, 01:04:07 PM
I'm happy to announce that Hatington Transport is now officially serving the greater Hatington area with passgener coach service!


On a more technical note: I get lots and lots of desyncs, which is annoying. I don't have the best internet connection, so that might be the problem... Which has the nice side effect that I now know the IP by heart :D
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 15, 2012, 01:36:20 PM
Hmm - may I ask, what are you doing immediately before the desync? One thing that can cause desyncs is if your client is slow, which will be exacerbated by having a large map. How much RAM do you have?
Title: Re: New Simutrans-Experimental server
Post by: rsdworker on January 15, 2012, 01:40:02 PM
james - how i do chat?
Title: Re: New Simutrans-Experimental server
Post by: ojii on January 15, 2012, 01:42:47 PM
Hmm - may I ask, what are you doing immediately before the desync? One thing that can cause desyncs is if your client is slow, which will be exacerbated by having a large map. How much RAM do you have?

I have 16GB of RAM.

There's no clear pattern. It can be that I'm just starting at a part of the map (no moving the viewport or anything) and get a desync...
Title: Re: New Simutrans-Experimental server
Post by: ӔO on January 15, 2012, 01:46:07 PM
james - how i do chat?
go to your message window
beside 'options' is a field where you can put text in.
Title: Re: New Simutrans-Experimental server
Post by: Ashley on January 15, 2012, 01:48:39 PM
Timothy - hmm, I am not sure why it is not announcing. Can I check that I have the various settings right in simversion.h? They are reproduced below:

Code: [Select]
/*********************** Settings related to network games ********************/

/* Server to announce status to */
#define ANNOUNCE_SERVER "servers.experimental.simutrans.org:80"

/* Relative URL of the announce function on server */
#define ANNOUNCE_URL "/announce"

/* Relative URL of the list function on server */
#define ANNOUNCE_LIST_URL "/list?format=csv"

/* Name of file to save server listing to temporarily while downloading list */
#define SERVER_LIST_FILE "serverlist.csv"

#endif

Are these correct?

Those settings are correct, yes. If you increase the debug level the game should print some announce-related logging information which may help you work out why it isn't working.

Edit: Also, I am having trouble using a non-root user to run Simutrans - I have set up a "simd" user, but when I run Simutrans with this user, it cannot accept any incoming connexions. Do you have any idea what to do about this?

This is on Debian Linux? Or some other platform? If you're trying to listen on a port below 1024 then only the root user can open a listen socket, however the Simutrans default is 13353 so you ought not to have an issue there. Check the output of the netstat -an command, grep for port 13353 and see which IP addresses your server is actually listening on. Remember for the announce server to register your game you need to be announcing the same IP address as the announce request comes from (or announce a DNS name which resolves to the address the announce request comes from).

How exactly are you running Simutrans as the alternative user? What do the logs say? Did you remember to chown the directory the game is running in so that that user can write there? Simutrans writes a number of temporary files during the course of a client connecting (since its serialisation model is file-oriented, it tends to just write a gameinfo/save file to disk and then transfer that file to the client, which is kind of dumb but isn't something I have the time to fix right now). My guess is this is likely to be your problem.
Title: Re: New Simutrans-Experimental server
Post by: ӔO on January 15, 2012, 02:02:06 PM
server seems to be down.
Title: Re: New Simutrans-Experimental server
Post by: rsdworker on January 15, 2012, 02:06:21 PM
server seems to be down.
yeah its looks like it
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 15, 2012, 02:23:44 PM
Hmm, no, seems to be running. Are you having trouble connecting?
Title: Re: New Simutrans-Experimental server
Post by: rsdworker on January 15, 2012, 02:25:36 PM
Hmm, no, seems to be running. Are you having trouble connecting?
yeah i am trying to connect

edit: i am connected now and there saving game in proccess
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 15, 2012, 02:27:02 PM
Hmm, odd - I didn't notice any interruption here.
Title: Re: New Simutrans-Experimental server
Post by: rsdworker on January 15, 2012, 02:28:11 PM
Hmm, odd - I didn't notice any interruption here.
the save is too frequently - its may be cause because its freezes client

after i clicked depot and tried to sell train - didn't vanish instantly - its still there :(
Title: Re: New Simutrans-Experimental server
Post by: ӔO on January 15, 2012, 02:36:02 PM
yeah, it is a bit sluggish when assembling, copying and starting convoys
Title: Re: New Simutrans-Experimental server
Post by: rsdworker on January 15, 2012, 02:36:37 PM
yeah, it is a bit sluggish when assembling, copying and starting convoys
plus building tracks or things - its takes while
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 15, 2012, 03:09:52 PM
Oddly, I find it quite responsive, although it does desync from time to time without user interaction. What ping are you each getting to the server?

I think that I shall have to restart the server having adjusted the map to 15fps, down from 30fps.
Title: Re: New Simutrans-Experimental server
Post by: rsdworker on January 15, 2012, 03:16:51 PM
Oddly, I find it quite responsive, although it does desync from time to time without user interaction. What ping are you each getting to the server?

I think that I shall have to restart the server having adjusted the map to 15fps, down from 30fps.
good idea

edit - i been disconnected again :(
Title: Re: New Simutrans-Experimental server
Post by: ӔO on January 15, 2012, 03:37:08 PM
I am getting around 113ms ping time.
Title: Re: New Simutrans-Experimental server
Post by: ojii on January 15, 2012, 03:48:57 PM
What ping are you each getting to the server?

rtt min/avg/max/mdev = 38.711/53.552/87.265/19.684 ms
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 15, 2012, 03:56:03 PM
Respectable pings (although I get about 37-47ms). I shall have to reduce the FPS when I get time.
Title: Re: New Simutrans-Experimental server
Post by: rsdworker on January 15, 2012, 03:57:01 PM
i had desyc trying to build bus line

i think around 200 ping because mine is fast
Title: Re: New Simutrans-Experimental server
Post by: Guvnor on January 15, 2012, 04:05:58 PM
Where do I go to see ping times in game?  I get no reply when pinging the IP outright.
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 15, 2012, 05:29:06 PM
Hmm - you should be able to ping it - I just tried, and got 14-15ms.
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 15, 2012, 05:58:43 PM
I have just restarted the server with a lower frames per second setting - can people reconnect and let me know whether it desyncs any less often? Thank you!

Edit: For reference, whoever is connected now (IP 81.62.xx.xxx), is pinging at about 37ms from the server.
Title: Re: New Simutrans-Experimental server
Post by: ӔO on January 15, 2012, 06:07:28 PM
I think I see one way to cause a desync and it has to do with copying/starting a lot of convoys at once by holding down the space key.
Title: Re: New Simutrans-Experimental server
Post by: ojii on January 15, 2012, 06:21:09 PM
Edit: For reference, whoever is connected now (IP 81.62.xx.xxx), is pinging at about 37ms from the server.

That's me!


Anyway, how do I build a public stop in Sim-Exp? I want to build a ship line to connect Willowingcaster with Curlden (to connect to give AEO the possiblity to connect to my network).
Title: Re: New Simutrans-Experimental server
Post by: ӔO on January 15, 2012, 06:24:42 PM
That's me!


Anyway, how do I build a public stop in Sim-Exp? I want to build a ship line to connect Willowingcaster with Curlden (to connect to give AEO the possiblity to connect to my network).

in ex, you connect networks by overlapping your station catchment area onto one of my stations.

btw, I was thinking of building down to nutworth, for now.
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 15, 2012, 06:41:07 PM
Oiji - hello! In Simutrans-Experimental, unless set otherwise in simuconf.tab, only the public player builds public stops. Instead of building public stops, you need to give access to the player(s) whom you wish to use your stops by clicking the appropriate checkbox in the players window.

As AEO pointed out, it is also possible for passengers (but not mail or goods) to walk between two nearby but not immediately adjacent stations, but, just as in real life, this is not ideal, as it adds transfer time. It will work, however, where the other transport company will not let you connect to its stations.

How's stability now, incidentally - any more desyncs?
Title: Re: New Simutrans-Experimental server
Post by: ӔO on January 15, 2012, 06:43:55 PM
I seem to be desynchronizing more often now.
Title: Re: New Simutrans-Experimental server
Post by: rsdworker on January 15, 2012, 07:16:14 PM
same as me - only when clients connect and disconnect - causes things not synced - example i get faster then when someone leaves game and game saves it then person leaves then i tried to send buses out its kept clicking till they are gone
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 15, 2012, 08:01:34 PM
Rsdworker - can you elaborate a little, since I am slightly confused - do you mean that you tend to desync just after somebody else has connected or disconnected? Or do you mean something else?

AEO - hmm, that's odd - all that I did was reduce the FPS, which should make things better. Can you try adjusting the additional_client_frames_behind in your simuconf.tab file (this is not saved with a game) to see whether changing this setting makes any difference?

Edit: Incidentally, some resource usage statistics, for those interested in resource consumption on this server. This instance is recorded as having started at 1755 and been running for 40:16 minutes. It is occypying 32.4% of the CPU (at the time of the check) and 16.8% of the 3Gb of RAM allocated to this virtual server.
Title: Re: New Simutrans-Experimental server
Post by: rsdworker on January 15, 2012, 08:03:49 PM
Rsdworker - can you elaborate a little, since I am slightly confused - do you mean that you tend to desync just after somebody else has connected or disconnected? Or do you mean something else?

AEO - hmm, that's odd - all that I did was reduce the FPS, which should make things better. Can you try adjusting the additional_client_frames_behind in your simuconf.tab file (this is not saved with a game) to see whether changing this setting makes any difference?

yes its occurs when clients connecting on and off - its causes desync for example client A - aeo on then other goes off - i was trying make bus line but took long time to send alot of buses
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 15, 2012, 08:22:11 PM
Hmm - very odd. Not quite sure what might be causing this.

Edit: Does this ever happen in Standard?
Title: Re: New Simutrans-Experimental server
Post by: rsdworker on January 15, 2012, 08:25:58 PM
Hmm - very odd. Not quite sure what might be causing this.
i had good look - the simloops was 1.6 when i tried to send out the buses
Title: Re: New Simutrans-Experimental server
Post by: ӔO on January 15, 2012, 08:29:47 PM
okay, I've restarted my game client with additional_client_frames_behind=8
Title: Re: New Simutrans-Experimental server
Post by: prissi on January 15, 2012, 08:34:41 PM
Different locations of trees => different building costs for ways => trees matters.

I did not say no trees, but after setting up the world, place trees manually. And set the tree per tile parameter to 2 or 3. That will greatly reduce map size, desyncs at season changes.

And with a map running at 80% single thread load, any more complex action is bound to cause a desync. For debugging desyncs, a smaller map which allows for faster reconnect seems more important. For the start try 512x512. There is still a lot that can be done on it.
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 15, 2012, 08:55:42 PM
AEO,

do let me know how you get on with the different settings.

Prissi,

hmm - why would the trees have different locations on the server/client? Are the tree data not part of the saved game, such that they synchronise automatically? Running a brief check with two clients simultaneously connected to the server, and testing a sample area, all the trees were in exactly the same place on both clients.

Also, it seems to me somewhat unlikely that the difference between players' profits would cause desyncs more than very, very rarely. To cause a desync, something would have to call a random number generator more often on the client than the server or vice versa (or possibly call it from different places). There is no direct connexion between a player's bank balance and any calls to random number generators. A desync is only likely to happen if building a particular thing would exceed a player's credit limit on the client but not the server (or vice versa), so the thing is built on the client and not the server (or vice versa) and the presence of that thing that is built in turn causes a desync.

As to CPU load - happily, this server is running at about 33% so far, so hopefully the 80% will not be reached for a while; but it is useful to know the limits.
Title: Re: New Simutrans-Experimental server
Post by: ӔO on January 15, 2012, 09:09:42 PM
I'm using additional_client_frames_behind=16 and still getting desynced.
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 15, 2012, 09:30:27 PM
Does it seem to make any difference, or is it equal with any setting? Are you getting desynced at particular times or randomly?
Title: Re: New Simutrans-Experimental server
Post by: ӔO on January 15, 2012, 10:05:03 PM
server seems to have crashed?
what is interesting is that the winxp client just closed down entirely, while the win7 client desynced.
the winxp client also never received your last message.
Title: Re: New Simutrans-Experimental server
Post by: prissi on January 15, 2012, 10:38:04 PM
@AEO:
How much is your CPU usage per core? If the server cannot follow certain commands, you time will be fast and you will be disconnect due command from the past.

That was why I suggested to have a map with a little less load.
Title: Re: New Simutrans-Experimental server
Post by: ӔO on January 15, 2012, 10:45:35 PM
@prissi
from the local save, I get around 25~35% cpu usage and around 380MB memory usage.
I'm not sure what I had in online mode.

edit: actually, the local save seems to have randomly crashed as well.

edit2: seems to happen at around 2:18~2:50 mark consistently, so it might have something to do with a way usage permission and convoys trying to use it.
Title: Re: New Simutrans-Experimental server
Post by: prissi on January 15, 2012, 11:03:53 PM
Well, random crashes that concludes my extremely small experience with actually playing sim-exp. But that was half a year ago.
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 15, 2012, 11:26:44 PM
I am looking into the crash issue presently.
Title: Re: New Simutrans-Experimental server
Post by: prissi on January 15, 2012, 11:39:43 PM
Use a tool like valgrind, which really help to iron out all issues with unitialized values and things depending on them. Such will be a quite sure case of desync or crashes. (Work on Linux. But just install unbuntu in a virtual machine like vmplayer works great.)
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 15, 2012, 11:41:37 PM
The crash that I was able to reproduce was a stack overflow, which I caught in MSVC++, associated with the generation of pedestrian graphics when passengers arriving at a stop are able to walk to their next transfer. I am looking into limiting the number of pedestrian graphics that can be generated per sync step to deal with this issue.

Edit: As should be apparent, it will not be possible to restart the server until I fix the bug and release the next (bugfix only) version of Simutrans-Experimental. In the meantime, here (http://simutrans-germany.com/files/upload/server13353-network.sve) is a copy of the saved game for the curious.
Title: Re: New Simutrans-Experimental server
Post by: sdog on January 16, 2012, 12:01:01 AM
SDog - actually, the trees don't seem to be causing much difficulty so far. For a high-specification map, a 45 second load time is fairly reasonable, I think. We shall have to see with experience whether it is unserviceable in play or not, but I really do want to test the boundaries of what can be done with high specification maps on a high specification server.
i don't doubt that your server can easily handle the large map, the bottleneck is the transfer. Especially with overseas connections.
Also keep an eye on the transfered data, so you don't run into limitations there. If you had three players connected on average and a desync every 5 minutes on average you have about 100 MB a hour, just for sending savegames after resets. (question to the devs: simutrans is transfering the whole game, or is it saved locally and just the differences transfered, similar to rsync?) That could easily get you some 60GB+ per month. With some contracts they have you pay through your nose if you exceed some limits.

trees had quite an impact on performance for me, much more than anything else in the game. but that might as well be due to the actual drawing, which doesn't matter for the server.
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 16, 2012, 12:13:30 AM
Sdog,

thank you for the thoughts. The server that I have has unlimited monthly bandwidth, happily, so that should not be an issue. (To answer your question, it transfers the whole game, albeit compressed in bz2 format).

With trees, I changed forsest.tab in Pak128.Britain some time ago to reduce the tree density and forest prepondernece to get a good smattering of trees but retain acceptable performance, as tree heavy maps caused me performance troubles, too.
Title: Re: New Simutrans-Experimental server
Post by: sdog on January 16, 2012, 12:19:33 AM
the more i think about stuff in simutrans, the more it hurts me that i still can't get along with c(++). (i really have to learn it, also as being good at c is a good insurance against unemployment.)
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 16, 2012, 12:21:39 AM
Do you know any computer programming languages?
Title: Re: New Simutrans-Experimental server
Post by: sdog on January 16, 2012, 12:31:39 AM
a few, but all are those easy 'pedestrian' languages, that do most for you, or are for people who are spoiled for programming. (fortran, perl, scripting, half a dozen used once and forgotten)
With structs, allocation, overloading, pointers and all object oriented i don't know the least bit. What's worth, i don't have an idea of theoretical computer science, which i am always told is the first step before even looking at a line of code.

What makes C++ almost impossible for me to read are the many casts.
To my very great shame, i was recently to stupid to use gsl (gnu scientific libraries) in perl for numerical integration. I coded a fortran 90 wrapper for 1980 fortran 77 subroutines and piped my data to that program :-(
In physics it doesn't matter, if it takes 3 seconds with that frankenstein, since i run the data through it once, and then the program is archived.

is Kunth still the best approach, to get into C?


ps.: i have the feeling i hijacked your thread (again!), would you like to split it off and move it to the off-topic discussions?
Title: Re: New Simutrans-Experimental server
Post by: ojii on January 16, 2012, 12:33:36 AM
a few, but all are those easy 'pedestrian' languages, that do most for you, or are for people who are spoiled for programming. (fortran, perl, scripting, half a dozen used once and forgotten)
With structs, allocation, overloading, pointers and all object oriented i don't know the least bit. What's worth, i don't have an idea of theoretical computer science, which i am always told is the first step before even looking at a line of code.

What makes C++ almost impossible for me to read are the many casts.
To my very great shame, i was recently to stupid to use gsl (gnu scientific libraries) in perl for numerical integration. I coded a fortran 90 wrapper for 1980 fortran 77 subroutines and piped my data to that program :-(
In physics it doesn't matter, if it takes 3 seconds with that frankenstein, since i run the data through it once, and then the program is archived.

is Kunth still the best approach, to get into C?


ps.: i have the feeling i hijacked your thread (again!), would you like to split it off and move it to the off-topic discussions?

C is NOT the same as C++, if you want to code on simutrans, I recommend you look into C++ directly.
Title: Re: New Simutrans-Experimental server
Post by: sdog on January 16, 2012, 12:36:40 AM
C is NOT the same as C++, if you want to code on simutrans, I recommend you look into C++ directly.
oh, i just used C to unspecifically address the language family, but had almost exclusively C++ in mind.
Title: Re: New Simutrans-Experimental server
Post by: Ashley on January 16, 2012, 12:49:50 AM
C is NOT the same as C++, if you want to code on simutrans, I recommend you look into C++ directly.

Heh, from what I've seen some parts of the code base are C++ only insofar as they have a .cc file extension :P

James - who're you buying this hosting from? I'd be interested to know where you're getting it so cheaply.
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 16, 2012, 12:59:24 AM
I get my hosting from Heart Internet (http://www.heartinternet.co.uk/). The price efficiency comes from the ability to customise the specifications: I save a lot of money by specifying the lowest amount of hard drive space (50Gb), and only one processor core (as Simutrans is single threaded), which means that I am able to get the server for £30/month including VAT.

Edit: Note that I have seen lots of people complaining about them failing to cancel services on request, although not so many about poor quality of service, so I took the precaution of using a debit card that was going to expire in April to pay.

Edit 2: Incidentally, the problem with the server game appears to be an infinite loop when passengers unload at Crowminster to transfer to/from the dock by foot - for some reason that I have not yet been able to diagnose, the passengers seem to circle endlessly between the two, which eventually creates a stack overflow.
Title: Re: New Simutrans-Experimental server
Post by: ӔO on January 16, 2012, 01:10:17 AM
while we are slightly off topic, I figure I might as well mention for james to check out the infrastructure costs for his freight company.
it's pretty close to 1/3 of my pax company no where near as profitable :/

I think there might be a need for road freight station with half the capacity and maintenance.
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 16, 2012, 01:41:08 AM
Hmm - my infrastructure maintenance is 1,090c to your 22,607c - are you sure? My margin is very low because my revenue is presently tiny.
Title: Re: New Simutrans-Experimental server
Post by: ӔO on January 16, 2012, 01:57:40 AM
My mistake, I actually meant the "next month fixed cost" field. The "Inf. maintenance" field is just "fixed cost * (months in operation)".
my fixed cost is 3,380c while yours is 1090c

your operating profit is roughly equal to your inf. maintenance, which means you are losing approximately 1090c/mo. or exactly your fixed cost.
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 16, 2012, 02:13:30 AM
Hmm. Do you have any suggestions for an appropriate early loading bay?
Title: Re: New Simutrans-Experimental server
Post by: ӔO on January 16, 2012, 02:34:54 AM
Hmm, well, I was thinking there was a lack of something in between.
There's the staging post, that is quite good for dropping off goods, due to its low cost and capacity (12).

Then there is the cargo bay that holds 128 and the station extensions that also hold 128.
I'm thinking the cargo bay can be halved in cost, maintenance and capacity.
Maybe knock off 1/4 of the cost and maintenance for the freight station extensions.
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 16, 2012, 02:40:56 AM
I think that I have fixed the crash bug, which fix is pushed to my 10.x branch; could those who are able to compile the code re-test in single player mode? I shall not be able to restart the server until I have put this fix into a new release, and I might as well take the opportunity to fix another bug or two that has come to my notice in the meantime, but hopefully I shall be able to get the server running again in a day or two.

Thank you for all your help in testing!

(And AEO - do you have any ideas for a real life prototype for such a thing?)
Title: Re: New Simutrans-Experimental server
Post by: ӔO on January 16, 2012, 03:01:30 AM
I highly suspect that everything was simply loaded by manual labour.
In the case of coal, it seems the wagon would tip over backwards (like a dumper truck tipping its dumper) so that people could shovel coal into the back.

I think you could simply recycle the base level goods siding from rail.
Title: Re: New Simutrans-Experimental server
Post by: rsdworker on January 16, 2012, 05:06:25 PM
Hmm. Do you have any suggestions for an appropriate early loading bay?
the early days - its was just stageposts for loading and unloading things - more simllar to modern street loading bay
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 16, 2012, 05:16:22 PM
Perhaps the early staging post can be like the rail siding with some boxes and other gubbins piled up nearby?
Title: Re: New Simutrans-Experimental server
Post by: ojii on January 16, 2012, 09:20:39 PM
I think that I have fixed the crash bug, which fix is pushed to my 10.x branch; could those who are able to compile the code re-test in single player mode? I shall not be able to restart the server until I have put this fix into a new release, and I might as well take the opportunity to fix another bug or two that has come to my notice in the meantime, but hopefully I shall be able to get the server running again in a day or two.

Thank you for all your help in testing!

(And AEO - do you have any ideas for a real life prototype for such a thing?)

Could you link to that branch?
Title: Re: New Simutrans-Experimental server
Post by: prissi on January 16, 2012, 09:52:39 PM
About the earlier comments on how you selected the smallest server. YOu seemed to have selected a virtual server. The CPU you get on a virtual server is changing. Last year a group in the german forum were using a virtual server and run into ocasional troubles with large maps; even though most of the time everything was fine.
Title: Re: New Simutrans-Experimental server
Post by: rsdworker on January 16, 2012, 09:59:06 PM
About the earlier comments on how you selected the smallest server. YOu seemed to have selected a virtual server. The CPU you get on a virtual server is changing. Last year a group in the german forum were using a virtual server and run into ocasional troubles with large maps; even though most of the time everything was fine.
sounds that what i was experincing - maybe needs a better server idk
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 16, 2012, 11:07:53 PM
The branch is here (https://github.com/jamespetts/simutrans-experimental/tree/10.x). It is indeed a virtual server that I am using - a dedicated server would not be adequate value for money. What exactly do you mean when you refer to the CPU on a virtual server changing; and how does that affect Simutrans?

Out of interest, Sdog conducted some very useful tests and found that he was getting far more desyncs in Windows XP than in Windows 7 or that dustNbone was getting in Linux, the latter of which systems were (aside from the crash, which I think that I have now fixed) rather stable.
Title: Re: New Simutrans-Experimental server
Post by: sdog on January 16, 2012, 11:13:59 PM
"Sdog conducted some very useful tests and found that he was getting far more desyncs in Windows XP than in Windows 7"

you must have confused me with someone else, i never managed to get simutrans running on windows. (haven't tried hard though)
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 16, 2012, 11:25:36 PM
Oops - sorry! Must have been AEO...
Title: Re: New Simutrans-Experimental server
Post by: ojii on January 16, 2012, 11:59:19 PM
I think that I have fixed the crash bug, which fix is pushed to my 10.x branch; could those who are able to compile the code re-test in single player mode? I shall not be able to restart the server until I have put this fix into a new release, and I might as well take the opportunity to fix another bug or two that has come to my notice in the meantime, but hopefully I shall be able to get the server running again in a day or two.

Thank you for all your help in testing!

(And AEO - do you have any ideas for a real life prototype for such a thing?)

re-test what exactly? it compiles fine and everything. but what exactly should we look for?
Title: Re: New Simutrans-Experimental server
Post by: ӔO on January 17, 2012, 01:31:56 AM
Oops - sorry! Must have been AEO...
yep, that was me.

both of my machines should be more than adequate to run large maps of simutrans ex, so it is a bit of a mystery to me.
Title: Re: New Simutrans-Experimental server
Post by: dustNbone on January 17, 2012, 01:40:55 AM
I compiled the 10.x branch and ran the savegame from the server game, seems to work normally.  Looks like it was my boat link that killed the server, I've never really played with the passengers transferring on foot feature, but it seems to get the job done.

Also, is the "Make Stop Public" option gone?  I think I get the access rights options but I did like the idea of being able to just make a hub public without allowing access to my entire network, which is what I think the access options in the player menu do.  Might be missing something though, not actually 100% sure on how it all works. 
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 17, 2012, 09:41:57 AM
Ojii - the idea was to test for crashes or any weird side effects of the fix. Thank you for compiling it, though, and thank you to dustNbone for testing also.

AEO - I wonder whether the Windows XP problems come from the method in which the platform's previous inability to run Simutrans when compiled with IPv6 capability was fixed?

dustNbone - the "make stop public" feature was set in 10.4 to be only available to the public player if a particular option in simuconf.tab was selected, which has been added to the simuconf.tab for Pak128.Britain-Ex 0.8.3. This is because it is not realistic for a company to be able to pass on the costs of its infrastructure maintenance to the government with impunity. The access function is intended to have the same ultimate effect. If you want a player to be able to access part of your network, you should generally discuss this with the player, either in the in-game chat or in the forum. You and the other player should agree the conditions of the access, including what part(s) of the network should be accessed. If the other player breaches the agreement, then you can withdraw access (with or without a warning, at your discretion). This, I think, is a far more satisfactory system than the "make public" tool (which is still available to the public player).
Title: Re: New Simutrans-Experimental server
Post by: prissi on January 17, 2012, 12:19:49 PM
The IPv6 code just affects the initialisation. However, if you have set up a IPv6 tunnel on XP, the traffic will try to use it prefrably. MAybe the windows 7 did not use it. Since a tunnel is overhead and need extrat latency, it could cause more freuqent desyncs. OTher than that there are not really much OS dependent code that should relate to desyncs. Maybe sound support makes a difference between both.
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 20, 2012, 01:51:05 AM
This server is now back up, running Simutrans-Experimental 10.6. See here (http://forum.simutrans.com/index.php?topic=9006.new#new) for a download link.
Title: Re: New Simutrans-Experimental server
Post by: Milko on January 20, 2012, 06:35:13 PM
Hello

I'm playng as "Italy transport".

 :)

Giuseppe
Title: Re: New Simutrans-Experimental server
Post by: ӔO on January 21, 2012, 03:51:55 AM
there seems to have been some sort of fatal error with the game.
the server itself is up, but it looks like the game has crashed.

it says something about
Code: [Select]
minivec_tpl<T>::[]
index out of bounds: 7 not 0.1
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 21, 2012, 02:52:43 PM
AEO,

thank you for your report: I have found and fixed the problem, and the fix is on Github. I hope to restart the server with the new version soon. Sorry for the trouble!
Title: Re: New Simutrans-Experimental server
Post by: Ashley on January 21, 2012, 02:53:53 PM
Heh, definitely best for you to be doing this James, makes it easier to debug when it's your own server ;)
Title: Re: New Simutrans-Experimental server
Post by: ӔO on January 21, 2012, 03:00:54 PM
@james
no problem. I rather enjoy finding the bugs :)
Title: Re: New Simutrans-Experimental server
Post by: Milko on January 21, 2012, 03:55:00 PM
Hello

When I place a structure or when I buy / sell a vehicle I have a problem.
 After giving the command the server executes it after several seconds, the delay between the command and its execution is all. To build complex structures, which require you to place a piece after another, the wait is a lot. This delay is correct?

Giuseppe
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 21, 2012, 04:00:24 PM
The delay is caused by your local game running more slowly than the server. This is not specific to Experimental; have you ever tried playing Standard over the internet?
Title: Re: New Simutrans-Experimental server
Post by: Milko on January 21, 2012, 04:35:15 PM
Hello

I have never used the standard version through internet, so I can not tell if the problem exists in the standard version or whether it is less marked.
 Indeed it seems that more delays are slower PCs (I have two PCs at home and I will try to investigate better) but I am surprised that even on an Intel PC T2050 1,60Ghz the problem is so thick ...
When the server returns I will try to work with my wife's PC (Intel i7).
 I will tell my wife not to work when I use the game  ;) .

Giuseppe
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 21, 2012, 05:05:36 PM
I will tell my wife not to work when I use the game  ;) .

Giuseppe

Now, here's a man with the right priorities!
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 21, 2012, 06:06:54 PM
The server is back up, now running 10.7 with the bug that brought it down before fixed. Happy playing!
Title: Re: New Simutrans-Experimental server
Post by: Milko on January 21, 2012, 06:35:43 PM
Hello

With Intel i7 I have no waiting times. :)

But I have frequent desync (when I'm setting lines and trips), unable to reproduce, I am testing...

Now it's ok.

Giuseppe
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 21, 2012, 07:34:33 PM
Thank you for the testing! Most helpful.
Title: Re: New Simutrans-Experimental server
Post by: ӔO on January 21, 2012, 07:51:13 PM
ah, the new cannot delete public ways can be a bit of a problem
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 21, 2012, 07:53:56 PM
Hmm - I implemented it because people were deleting public roads and leaving cities cut off. It's more realistic this way, I think: if you want the public player (in this game, me) to move roads, post in this thread.

Edit: Note that this does not apply to city roads.
Title: Re: New Simutrans-Experimental server
Post by: AP on January 21, 2012, 08:03:28 PM
Having a bit of trouble playing; the server seems to spend a lot of time saving and loading the map, compared to actually *playing* it, and seems fond of kicking me out for no known reason...
Title: Re: New Simutrans-Experimental server
Post by: Milko on January 21, 2012, 08:05:28 PM
Hello

Now we are 4 players connected, but I think more players = more desyncs....

PS the game don't remember the server address, is it correct?

- Server off line? -

Giuseppe
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 21, 2012, 08:13:22 PM
It seems to have gone down briefly for reasons not yet entirely clear, but I have restarted it now.
Title: Re: New Simutrans-Experimental server
Post by: AP on January 21, 2012, 08:22:06 PM
I just get "server did not respond" now when I try and join @ 46.32.231.222
Title: Re: New Simutrans-Experimental server
Post by: ӔO on January 21, 2012, 08:23:37 PM
Hmm - I implemented it because people were deleting public roads and leaving cities cut off. It's more realistic this way, I think: if you want the public player (in this game, me) to move roads, post in this thread.

Edit: Note that this does not apply to city roads.

ah, okay, I thought it was all roads.
as long as city roads can be deleted it is okay. otherwise it becomes quite problematic trying to build the station near the city center.

seems like I couldn't delete the road due to desync.
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 21, 2012, 08:23:47 PM
Hmm, can you try again? It works for me, and reports being up from the command line. The IP is: 46.32.231.222
Title: Re: New Simutrans-Experimental server
Post by: ӔO on January 21, 2012, 08:27:43 PM
The server seems to be on and off.

I see a general trend of more players = more desyncs


and I think a player can still delete public roads by over writing it with their own road
seems not
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 21, 2012, 08:33:55 PM
It's still reported as being up on the command line...

I shall look into the issue about over-writing with other roads, though. Thank you for spotting that.

Edit: I can't reproduce the over-writing roads issue. Can you elaborate?
Title: Re: New Simutrans-Experimental server
Post by: AP on January 21, 2012, 08:48:03 PM
I can connect now, it tells me "2 clients connected". There's a lag between building something and it appearing of 2-3seconds. It also has disconnected me twice, roughly 3-4 minutes after joining each time.
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 21, 2012, 08:54:32 PM
AP - lags between commands being sent and executed is usually as a result of your client running more slowly than the server. I am not sure, but I suspect that this might also cause desyncs eventually. May I ask - what are the specifications of the computer that you use to connect?
Title: Re: New Simutrans-Experimental server
Post by: AP on January 21, 2012, 08:56:13 PM
Lenovo W500
Intel Core2Duo 2.53Ghz
3072mb Ram
ATI mobility radeon hd 3650
Windows 7
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 21, 2012, 09:03:07 PM
Hmm. A decent enough specification, although not on the same level as the i7s that people are reporting work well. Is it any better now?
Title: Re: New Simutrans-Experimental server
Post by: AP on January 21, 2012, 09:11:45 PM
Alas not. If anything (and it's subjective so I can't be sure), the lag increased.
Title: Re: New Simutrans-Experimental server
Post by: ӔO on January 21, 2012, 09:13:12 PM
hi james, I need some public player assistance in rerouting some roads on the south western side of the map.

it seems like I have the lowest specs, but both systems only get around 30~40% CPU usage.
intel C2D, 2.13Ghz
3GB RAM
AMD HD5770
WinXP

and

AMD E-350 1.6ghz
6GB RAM
AMD HD 6319 384MB
Win7
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 21, 2012, 09:26:08 PM
Do you find that your performance is acceptable, AEO?

AP and AEO - can you tell me your pings?
Title: Re: New Simutrans-Experimental server
Post by: ӔO on January 21, 2012, 09:29:15 PM
ping is 120ms average.

with up to 2 players playing, it seems to stutter, but with 4 it seems like I will desync. by stutter, I mean it is smooth sometimes and sluggish other times.
Title: Re: New Simutrans-Experimental server
Post by: AP on January 21, 2012, 09:54:57 PM
How do I get it to tell me my ping?
Title: Re: New Simutrans-Experimental server
Post by: ӔO on January 21, 2012, 10:07:12 PM
How do I get it to tell me my ping?
I just do it from windows command line.
win key+R -> cmd -> ping 46.32.231.222
Title: Re: New Simutrans-Experimental server
Post by: AP on January 21, 2012, 10:17:56 PM
Thanks. There may have been a time when I knew how to do that, but it's long forgotten!

Ping average 102ms
Title: Re: New Simutrans-Experimental server
Post by: rsdworker on January 21, 2012, 10:26:05 PM
mine is 33ms
Title: Re: New Simutrans-Experimental server
Post by: AP on January 21, 2012, 10:38:04 PM
Well, I'm done for the evening. In some respects it clearly works well, but certain elements - notably trying to do anything in-depot, were slow to the point of being unresponsive. AEO appeared to be having a more successful in-game conversation with James about it. I found my in-game messages had a lag on them - I'd type them concurrently with the discussion occurring, and they'd appear 4 or 5 minutes later. No idea if this was due to haivng 3/4 clients connected.
Title: Re: New Simutrans-Experimental server
Post by: rsdworker on January 21, 2012, 10:46:05 PM
Well, I'm done for the evening. In some respects it clearly works well, but certain elements - notably trying to do anything in-depot, were slow to the point of being unresponsive. AEO appeared to be having a more successful in-game conversation with James about it. I found my in-game messages had a lag on them - I'd type them concurrently with the discussion occurring, and they'd appear 4 or 5 minutes later. No idea if this was due to haivng 3/4 clients connected.
i seen that when i used it - i think i noted it this thread
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 22, 2012, 12:18:32 AM
Hmm - I didn't have any difficulty with chat lag being as much as 3-4 minutes. That will generally mean that your client is running the game slower than the server, so might indicate relatively low performance hardware; conversely, AEO doesn't have anything spectacular, so it's not clear why there should be such a performance difference between the systems. AEO - did you find that in-depot performance was acceptable?
Title: Re: New Simutrans-Experimental server
Post by: ӔO on January 22, 2012, 02:19:07 AM
To repeat what was said in-game, so other can see.

For me, it seems like the depot is the number one spot where I will desync. Purchasing new vehicles seems to be more problematic compared to assembling from available inventory.

convoy replacement is a guaranteed desync, however the command will be executed.

Generally, it seems that 4 clients or more will cause severe instability, while 2 will result in a slight delay between actions.
Title: Re: New Simutrans-Experimental server
Post by: rsdworker on January 22, 2012, 02:20:34 AM
To repeat what was said in-game, so other can see.

For me, it seems like the depot is the number one spot where I will desync. Purchasing new vehicles seems to be more problematic compared to assembling from available inventory.

convoy replacement is a guaranteed desync, however the command will be executed.

Generally, it seems that 4 clients or more will cause severe instability, while 2 will result in a slight delay between actions.
i have simllar experince - when i had 2 clients on  then i was happy building away things then suddenly 4 clients its started to deysnc
Title: Re: New Simutrans-Experimental server
Post by: ӔO on January 22, 2012, 07:16:09 AM
server might be down.
no client side crash this time.

it seems like people were trying to connect to the server, but couldn't.
Title: Re: New Simutrans-Experimental server
Post by: Milko on January 22, 2012, 11:05:43 AM
Hello

This is my experience last night.
 Delayexecutionofcommands:
 10/15 seconds on an Intel Core Duo 1.6 Ghz 1Mb RAM
 Instant on an Intel i7 4Mb RAM

 desync:
 Present (seems random) when working on lines and routes of the convoys.
 Increase as people connected.
 Yesterday, during a time when I was alone and I have not used the vehicle management, I was offline for about 15 minutes without interruption.

Giuseppe
Title: Re: New Simutrans-Experimental server
Post by: AP on January 22, 2012, 11:30:27 AM
Server seems to be down now.
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 22, 2012, 01:45:14 PM
I have restarted the server: apologies that it went down. I have also enabled logging and an automatic check as to whether the server is running every 5 minutes, and an automatic restart if it is not. That should keep it running!
Title: Re: New Simutrans-Experimental server
Post by: ӔO on January 22, 2012, 05:18:11 PM
did the server crash?
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 22, 2012, 05:27:58 PM
Yes, as did my local client, but I couldn't reproduce it when I opened the saved game on my local machine and ran it in a debugger - the crash must have been induced by something that somebody did (the logfile suggests that people were building railways), such that I can't reproduce it.

I have used the opportunity to test the automatic restarting script, which seems to be giving some trouble.
Title: Re: New Simutrans-Experimental server
Post by: ӔO on January 22, 2012, 05:31:25 PM
yeah, I thought it was a crash relating to a bug as well, since my client crashed right after I desynced.
The locally save game worked fine, so I was puzzled.
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 22, 2012, 05:46:14 PM
Right - I have restarted the server, but my cron jobs are not running for reasons that I do not understand; even a test cron job (syntax */1 * * * * echo "TESTING" >> /var/log/simutrans/test.log) fails.

Does anyone with any experience of cron jobs have any suggestions? I have checked that the cron deamon is running, and it appears to be:

Code: [Select]
root@438242:~# ps -ef|grep cron
root       687     1  0 17:16 ?        00:00:00 cron
root      1138   948  0 17:36 pts/0    00:00:00 grep --color=auto cron

The odd thing is that it seems to have run once the very first time that it was scheduled, and to have failed thereafter. Any suggestions as to cron would be much appreciated. In the meantime, apologies for the interruption to the server.
Title: Re: New Simutrans-Experimental server
Post by: AP on January 22, 2012, 07:36:38 PM
Is it normal for the game to periodically save/load when playing a network game? Mine keeps doing it every few minutes, and it takes 30s or so. Not sure if it's linked to the other performance issues I've had with the game though.

I did have one other thought - the map has a lot of trees on it. When I was doing work on the UK maps, which were similarly large files, I was advised by someone (prissi?) that trees use memory/something which will decrease performance. Wondered if a 'bald' map might be worth testing?
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 22, 2012, 07:45:22 PM
Hmm - I haven't noticed any loading and saving other than when players join. As to trees - this would affect memory consumption more than anything, and only a small proportion of the server's available memory (around 17%) is currently used.
Title: Re: New Simutrans-Experimental server
Post by: AP on January 22, 2012, 07:57:39 PM
I suppose with all the desyncing going on, that could have been what was causing it.
Title: Re: New Simutrans-Experimental server
Post by: Junna on January 22, 2012, 07:58:39 PM
Did it go down again?
Title: Re: New Simutrans-Experimental server
Post by: AP on January 22, 2012, 08:07:13 PM
I'm in, so I assume not.
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 23, 2012, 12:11:51 AM
Incidentally, some testing suggests that increasing additional_client_frames_behind in simuconf.tab to 3 or even 4 seems to reduce desyncs.
Title: Re: New Simutrans-Experimental server
Post by: AP on January 23, 2012, 10:22:37 AM
Don't know if anything changed overnight, but I had quite good performance/low lag last night with 2-3 clients connected, today, large lag in similar circumstances (30s between click and action occuring).
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 23, 2012, 11:21:57 AM
The lag may be due to network conditions - I haven't changed anything on the server itself, certainly. How are you doing for desyncs?
Title: Re: New Simutrans-Experimental server
Post by: AP on January 23, 2012, 11:27:28 AM
Could be. Only one desync today so far. I have manually disconnected / reconnected to try to reduce lag (variable success).
Title: Re: New Simutrans-Experimental server
Post by: Milko on January 23, 2012, 11:28:43 AM
Hello James

- delay in command executing -
The reduction in the number of frames per second set in the server allows to improve the response time to commands on the client?

Giuseppe
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 23, 2012, 11:35:59 AM
Ahh, no: it makes the delay in executing (slightly) worse, but reduces the chance of desyncs.
Title: Re: New Simutrans-Experimental server
Post by: AP on January 23, 2012, 01:58:57 PM
Right, I'm concluding I'm rubbish at Experimental, need more practice in private I think, having near bankrupt 3-companies!
Title: Re: New Simutrans-Experimental server
Post by: ash1794 on January 23, 2012, 04:55:27 PM
is this server down? cause i tried joinin and failed... "SERVER DID NOT RESPOND"..
Title: Re: New Simutrans-Experimental server
Post by: ӔO on January 23, 2012, 05:59:40 PM
yep, it seems to be down.
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 23, 2012, 09:14:02 PM
Hmm - it's not down for me...

Edit: Having checked, I think that my cron job for auto restarting the server has suddenly started working again!
Title: Re: New Simutrans-Experimental server
Post by: Milko on January 25, 2012, 09:31:54 PM
Hello

So things are looking better, I can stay connected for about 10 minutes before a desync, even when we are connected in two or three.

We must still work on desync, but now you can play decently. Thanks James.  :)

Edit: now we are connected in 6 players!!!

Giuseppe
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 25, 2012, 09:53:51 PM
Giuseppe - yes: I am player 6! As for desyncs, you can reduce these by increasing your additional_client_frames_behind setting.
Title: Re: New Simutrans-Experimental server
Post by: ӔO on January 26, 2012, 07:57:57 AM
james, when you have the time, I need to have roads fixed up between.
- Axemead, Catingfield
- Scarmead, Reinwich
- Reidwich
- Kingshampton
- Pointshore, Cheringfield
- Cheringfield, Workmere
- Mathorn (your sign is in the way)

thanks!
Title: Re: New Simutrans-Experimental server
Post by: Guvnor on January 26, 2012, 03:00:16 PM
I'm pretty sure the server's crashed.
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 26, 2012, 07:14:59 PM
Hmm, seems indeed to have crashed. Investigating...
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 26, 2012, 08:56:45 PM
The server is back now: the error that caused the crash was most peculiar, as was the fact that the way that I was able to get it working again worked (if any coding minded person is interested in helping to solve the mystery, a detailed explanation is available here (http://forum.simutrans.com/index.php?topic=9061.new#new)).

Sorry for the outage!
Title: Re: New Simutrans-Experimental server
Post by: Guvnor on January 26, 2012, 11:20:22 PM
James,

It crashed again.  I've attached a screen shot.  Notice the load bar behind.

Dave
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 26, 2012, 11:25:47 PM
Yes, I was online at the time - trying to track down the issue. It's back up again from an earlier saved game: sorry if anything has been lost.
Title: Re: New Simutrans-Experimental server
Post by: jamespetts on January 27, 2012, 12:26:58 AM
Hmm - the quickstone error (discussed here (http://forum.simutrans.com/index.php?topic=9061.new#new)) seems to be occurring too frequently to deal with effectively manually, so it will have to be fixed before we can go live again, I'm afraid. There's a workaround on my 10.x branch, but I don't have time to deploy to-night, so I'm afraid that AEO will have to take a brief break from world domination!
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on January 29, 2012, 12:17:33 AM
The server is now back online with 10.8, and is also now on the announce server (http://servers.experimental.simutrans.org) - thank you everyone for your patience.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on January 29, 2012, 12:40:27 AM
is it just me, or does it seem like the save/load takes longer now?
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on January 29, 2012, 12:46:18 AM
Hmm - I haven't noticed anything in particular.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: dustNbone on January 29, 2012, 01:19:43 AM
Seems about the same to me.  I would think it would get progressively larger as the game progresses.  Overall it actually seems a touch more responsive to me. Could just be weekend network conditions but wondering if maybe the bug was causing the server to chug away on some endless task.  Anyway seems to be working fine.  Except my client just segfaulted during a map reload after someone connected.  :)
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on January 29, 2012, 01:30:16 AM
Hmm - you don't happen to have a backtrace, do you?
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: dustNbone on January 29, 2012, 01:34:13 AM
No I forgot to copy it out of the terminal before I restarted the client, then it was past the scrollback buffer.  I'll try to remember next time.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on January 29, 2012, 01:36:04 AM
Thank you.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on January 29, 2012, 01:38:31 AM
okay, it's probably just my senses throwing me off.
so far, so good :)
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on January 29, 2012, 01:46:06 AM
I am noticing that it seems to take other clients longer to load: the time that I am paused after loading is longer than it was before, but the loading itself is not.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: TurfIt on January 29, 2012, 01:59:15 AM
When I connected previously to browse around, I noticed it taking a very long time to save/reload (this was the older 10.7). I was barely finished the save/reload cycle before it unpaused when someone joined. Not a super fast machine but an old Core2 @3.3GHz that was pegged for the duration. Changing my local saveformat to zipped instead of bzip2 was a huge help. I suggest that those temporary saves done on all clients should be zipped rather than bzip2 (or even binary), but leave the server with bzip2 for faster transfer - ~50% difference in file size.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: sdog on January 29, 2012, 02:01:25 AM
just a question, can the time-line time slowed down for a server? So that a year takes longer in real life time. However game time for vehicle movement stays the same. (eg ticks per month increases at the same rate).
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on January 29, 2012, 02:11:08 AM
TurfIt,

that's a very interesting idea that merits some consideration.

Sdog,

I am not quite sure what you mean here: I can change the ticks per month setting, but do you mean something different to this?
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: sdog on January 29, 2012, 02:21:57 AM
increasing ticks per month would keep the speed of the game, at which the vehicles operate, the same? But  would slow down the progression of years? (most non-pak.britain servers play with timeline off, as they reach the time where no vehicles are available anymore too quickly.)
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on January 29, 2012, 02:35:13 AM
I think we have done... around 30 years in 2.5 weeks?


edit:
James, I will need the roads at Dartbury, Perlshore, Catdon, Newham and Axingworth touched up.
at dartbury, your sign is on the road in one corner of the station.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on January 29, 2012, 10:44:44 AM
increasing ticks per month would keep the speed of the game, at which the vehicles operate, the same? But  would slow down the progression of years? (most non-pak.britain servers play with timeline off, as they reach the time where no vehicles are available anymore too quickly.)

Yes, that's the whole purpose of ticks per month - is it going too quickly, do people think? I can change it for the next game.

AEO - will have a look at those issues.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on January 29, 2012, 12:18:36 PM
I think you have stuck my curlden station.

I can't seem to get any command through right now, although I can see what is going on.


edit: well it did get stuck, but eventually unstuck itself.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: AP on January 29, 2012, 07:11:31 PM
I get he following on trying to connect to bridgewater-brunel.me.uk

FATAL ERROR:
quickstone <T>::quickstone_
tpKT(T * ,uint16
slot (16) already taken
PRESS ANY KEY

and then simutrans crashes.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on January 29, 2012, 07:30:27 PM
Hmm - I can't reproduce this. Are you sure that you are using the latest version 10.8?
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: AP on January 29, 2012, 08:47:37 PM
Oh ****, I hadn't seen the update. That fixed it. Map looks to be going well!
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on January 29, 2012, 09:52:12 PM
Excellent!
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on January 30, 2012, 01:44:03 AM
James, sorry to bother again, but I will need road works at Faringworth and Carmere

edit: and between kilcaster and burham
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: Junna on January 30, 2012, 05:59:06 AM
I need a realignment at Nutmere and Nutworth (should be pretty straight forward.)
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on January 30, 2012, 10:11:11 PM
I have done these, except for Nutworth, where I cannot find an issue - what did you want done there?
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: Junna on January 31, 2012, 02:07:37 AM
I have done these, except for Nutworth, where I cannot find an issue - what did you want done there?
The bridge should be four tiles (now three), and I'd like two track space on the non-city side of the station to allow the express tracks through (currently only one).
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on January 31, 2012, 08:12:24 AM
the server appears to have bugged out on a save/load
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: dustNbone on January 31, 2012, 09:03:29 AM
Yeah the save croaks my client too.  Terminal just outputs Segmentation Fault, didn't have time to run a debugger.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on January 31, 2012, 10:52:36 AM
Can confirm that the server seems to have gone down at 0751h this morning. Will have to investigate presently.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 03, 2012, 12:34:59 AM
The server is back up again, now running 10.9 - apologies for the delay in getting it restarted: have been very busy at work. You will need to upgrade to 10.9 to join.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on February 07, 2012, 02:09:18 AM
- There is a depot at saltington with ownership under the public player. For some reason, it wasn't deleted when the company there was liquidated.

- At Elton, a bridge is bugged out.
- At Parfield, could the at-grade crossing road be removed? It's quite painful to watch a train going 120km/h suddenly come to a halt because a 6km/h horse and carriage is going over the tracks.

- At Dartbury, there is a piece of road still left under the bridge. Many bridges had to be built, because the grade crossings were killing express trains.
- Same at Midingford, Pentham, Nightingborough, Newham, Ardingfield, Nightworth, Rainwich.

- Beamead has no road connection
- River at Innodon is disconnected
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: Milko on February 07, 2012, 09:28:55 PM
Hello

My company continues to lose money every year but my cash is 830M; does not add up ...

Giuseppe

Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on February 07, 2012, 09:54:05 PM
it is your interest rate.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: Milko on February 07, 2012, 10:27:53 PM

it is your interest rate.

Strange, unfortunately, have gone 10 days since my last connection and can not remember my last position. In memory, however, was convinced that the interests were not enough to offset losses.

   One question ... the game tells you how many players are connected, does anyone know if you know who are these players?

Thanks

Giuseppe

Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on February 07, 2012, 10:49:20 PM
early in the game, the interest rate is not much, but once you have a sizable account balance, the interest rate can keep you solvent, even if you are posting a loss.

I wish I knew which players are the ones connected too.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: Junna on February 08, 2012, 05:35:59 AM
Instability and constant desynchs have persisted for a few hours now. Someone connects - another drops.

(Also I keep crashing when someone else connects, and occasionally when I connect myself.)
 
Was stable for two hours. Desynchs a' plenty once more, however. Wonder if it has something to do with the replacer?
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: Milko on February 08, 2012, 09:19:48 AM
Hello

Last night I connected with you and AEO, but I was as an "observer" because I lost my password ...  :-[
I had a desync every 10 minutes (not bad), but of course I was not operating. We were in four..

Giuseppe
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 08, 2012, 11:03:06 AM
Thank you for the report as to the desyncs and crashes. I was logged on last night, and did notice this. I have not been able to reproduce the crashes, unfortunately: if anyone is able to find reliable steps to reproduce, I should be most grateful.

As to the desyncs, I tried tinkering with some of the network settings in simuconf.tab yesterday evening when the problems started. I set server_frames_ahead = 6 (previously 5) and server_frames_between_checks=240 (previously 320) and also server_frames_per_step=5 (previously 4).

I shall try setting server_frames_per_step=3 to see whether that helps.

Edit: Server is restarted with the modified setting. Please let me know whether that helps. Incidentally, desyncs have in the past been associated with network peak time, so I should be interested in whether anyone else can confirm this correlation or find any other correlations.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: prissi on February 08, 2012, 03:25:38 PM
With server_frames_between_checks=240 only every 10s desyncs are detected (if runnign at 24 fps). Thus the reason may have already been lost.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 08, 2012, 03:38:40 PM
Prissi,

thank you for the tip. Does this apply only to desyncs based on the random number generator status (i.e., general desyncs) or does this apply to all desyncs, including those related to the client running ahead of the server?
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on February 08, 2012, 07:34:26 PM
whatever happened, the desyncs are now even worse.
not only that, but there are constant client crashes upon save/load
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: Junna on February 08, 2012, 07:44:12 PM
whatever happened, the desyncs are now even worse.
not only that, but there are constant client crashes upon save/load

What AEO said. I can scarcely stay on for more than 20 seconds before it desyncs. And crashes occur on every other or third save/load.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on February 08, 2012, 08:00:59 PM
I can quite reliably cause a crash when using the convoy replacer. I don't know if it crashes the server, but the client sure does crash.

in the replacer using clear, creating a new convoy, then removing a vehicle will cause a crash. For me, I was accidentally selecting the midland 483 when I wanted the 1000. So when I went to remove the tender manually, it would crash, however it doesn't crash when the "clear" button is used.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 08, 2012, 09:55:57 PM
Hmm - interestingly, Simutrans-Experimental does not seem to be the only thing affected by this spate of desyncs: see here (http://forum.simutrans.com/index.php?topic=6547.new;topicseen#new).

I'm afraid that I am having trouble reproducing the replacer crash, however. I shall try to log in with a debug version to see whether I can reproduce the desyncs or the load/save crashes. Apologies that you have been having trouble.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: prissi on February 08, 2012, 10:01:52 PM
Frequent desync happen often when people connect with the wrong version. I built a version check based on GIT into your makefile for mingw. Are you using it?
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 08, 2012, 11:04:50 PM
Prissi,

I must confess, I did not remember that you had done that, and I do not think that I have implemented it - I suspect that I must have been very busy when you first mentioned it, and I must have forgotten when I had more time again.

However, I did ask the players which version that they were using, and both who were online were - correctly - using 10.9.

Might I ask with respect to the server_frames_per_step setting - which is more conservative: a higher or lower number?
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: prissi on February 08, 2012, 11:40:22 PM
No idea, was added by Dwachs, if memory servers me right.

But you should give releases a unique number, like the revision number from SVN. I posted code for you, but I do not find it (and my build computer was set up new, thus I do not have this kind of code.)

It was built around "git log --oneline" which gives a revision number in the beginng. Then this line was hashed to get a 32 values which could be used a unit server version ID string.

Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 08, 2012, 11:44:43 PM
At present, I simply "#define REVISION 109" for 10.9, etc., which should suffice for most purposes, although, for some reason that I have not had a chance to look into, the checking of this does not seem to work.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on February 09, 2012, 12:05:54 AM
aside

From Workmere to Woolham, there are two trains running. One is a direct express route service (Rail central | express - thurmouth to woolham), while the other goes around in a wide circle (Rail east branch - hatley (wide)). A lot of the passengers between these two points are taking the slower, longer route for some reason.
Is it because of frequency? Comfort is better in the express too.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 09, 2012, 12:13:47 AM
AEO,

it might be frequency, but it won't be comfort, since comfort is not taken into account in routing (it would be too computationally expensive to do so). It is waiting time that makes frequency affect passenger routing choices.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 09, 2012, 01:45:17 AM
Hmm - by my testing this evening, it looks as though we might have a simrand desync issue connected to private cars, which has not shown up previously, as there were hitherto very few private cars; the desyncs that we have been getting are now confirmed as simrand desyncs (rather than the server/client timing desyncs that we were getting in the early days), and connecting locally to a game (Sdog's 173) in which private cars are in abundance causes a near instantaneous simrand desync, whereas connecting to an archive of the server map from 1857 is stable.

This will take some investigating. If anybody in the meantime would like to look into the code to find the source of this problem, I should be extremely grateful. A pointer for those new to the network code: simrand desyncs occur when the random number generator (simrand - which is actually deterministic for any given seed, which seed is shared between the server and clients) is called a different number of times on the client and the server in the same number of steps. This, in turn, is caused by the code path that the program follows diverging somewhere on the client and server - normally, they are supposed to run exactly in parallel (apart from the GUI).

Deviations in the code path are caused principally by two different sorts of things: (1) variables not being loaded/saved properly (so that a value saved on the server is not transmitted to the client, or if it is lost or improperly adjusted after being transmitted to the client such that the variable is different on the client and server ultimately leading to a code path deviation); and (2) undefined behaviour in the code (reading beyond the end of an array, etc.) causing truly random rather than the desired controlled psudo-random behaviour. This sort of bug is notoriously difficult to find, because it is very difficult to pin down where it happens.

In this instance, I have confirmed that it appears in versions of Simutrans-Experimental between the current 10.x branch and 10.8 inclusive. The code for private car journeys, where I very strongly suspect that the problem lies, is found in simcity.cc. The code for actually moving the graphics on the map is elsewhere, but I suspect at present that this element is not problematic.

To test network server/client connectivity (which works well for checking simrand desyncs), run both a server and client on your local machine, and connect to the loopback interface (127.0.0.1) in the network connect dialogue.

I shall work on this (as a priority) when I get time, but if anyone were able to narrow it down, I should be very grateful. In the meantime, apologies most sincere for the disruption to your games.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on February 09, 2012, 03:23:52 PM
AEO,

it might be frequency, but it won't be comfort, since comfort is not taken into account in routing (it would be too computationally expensive to do so). It is waiting time that makes frequency affect passenger routing choices.
on closer observation, what seems to be happening is the express trains are too frequent and they do not wait for passengers to board before they turn around. The branch line on the other hand waits for a long time, which allows passengers to board it.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: prissi on February 09, 2012, 04:29:07 PM
Are you still using pointerhastables for sync_step or in any other location? Then this will almost inevitely lead to fast desyncs.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: Dwachs on February 09, 2012, 07:50:31 PM
Might I ask with respect to the server_frames_per_step setting - which is more conservative: a higher or lower number?
You mean server_sync_steps_between_checks ? The lower the number the more often the server sends sync information and the earlier a desync will be detected.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 09, 2012, 09:44:31 PM
A brief update on the desync issues: on a local test server, with the saved game loaded automatically when Simutrans-Expeimental starts, connecting to it does not desync. However, when loading a game manually, including that very same game re-saved manually, it desyncs almost instantly on connexion. This suggests to me an error in the load/save routine somewhere.

To answer Prissi's question - I vanquished the ptrhashtables from sync steps a long time ago, so it's not that, alas. Dwachs - no, I know what server_frames_between_checks is (I presume that that is what you meant by "server_sync_steps_between_checks"?); I was asking about this parameter:

Code: [Select]
# In network mode, there will be a fixed number of screen updates before a step.
# Reasonable values should result in 2-5 steps per second.
server_frames_per_step = 4
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 10, 2012, 12:09:39 AM
Right, I think that I have narrowed down the problem: it appears to have arisen first in version 10.8, and seems associated with the fix for the quickstone errors. It occurs only in saved games which, when loaded with version 10.7 or earlier, fail with quickstone errors. Any game that 10.7 can load without quickstone errors will, from my testing so far, work without desyncing, even when loaded in later versions.

I wonder, therefore, whether I have not fixed this issue fully. I think that it was Dwachs who spotted the initial problem - do you have any thoughts as to what I might have missed in fixing this that might cause desyncs?

Edit: I am suspecting that this problem might be related to this (http://forum.simutrans.com/index.php?topic=9219) crash: this is caused when, in the following code (which is largely the same as in Standard, apart from layout and the Experimental file version):

Code: [Select]
// will restore halthandle_t after loading
    if(file->get_version() > 110005)
    {
        if(file->is_saving())
        {
            uint16 halt_id = self.is_bound() ? self.get_id() : 0;
            file->rdwr_short(halt_id);
        }
        else
        {
            uint16 halt_id;
            file->rdwr_short(halt_id);
            self.set_id(halt_id);
            if(file->get_experimental_version() >= 10 || file->get_experimental_version() == 0)
            {
                self = halthandle_t(this, halt_id);
            }
            else
            {
                self = halthandle_t(this);
            }
        }
    }
    else
    {
        if (file->is_loading())
        {
            self = halthandle_t(this);
        }
    }

self = halthandle_t(this, halt_id); is called when halt_id == 0. The problem, then, seems to come from the fact that sometimes halts are saved with a zero id (which is explicitly permitted by the saving code: uint16 halt_id = self.is_bound() ? self.get_id() : 0;.

It would seem sensible for rdwr not to be called on a halt whose "self" halthandle is not bound; indeed, something has probably gone wrong to cause the "self" handle to be unbound in the first place (any thoughts on what that might be?), but perhaps the save routine should not do something that will inevitably cause a crash on loading in this case?

Any thoughts on this development much appreciated!

Edit 2: I wonder whether this problem might be related to this code to save waiting times:

Code: [Select]
for(short i = 0; i < max_catg_count_file; i ++)
        {
            if(file->is_saving())
            {
                uint16 halts_count;
                halts_count = waiting_times[i].get_count();
                file->rdwr_short(halts_count);
           
                inthashtable_iterator_tpl<uint16, waiting_time_set > iter(waiting_times[i]);

                halthandle_t halt;
                while(iter.next())
                {
                    uint16 id = iter.get_current_key();

                    if(file->get_experimental_version() >= 10)
                    {
                        file->rdwr_short(id);
                    }
                    else
                    {
                        halt.set_id(id);
                        koord save_koord = koord::invalid;
                        if(halt.is_bound())
                        {
                            save_koord = halt->get_basis_pos();
                        }
                        save_koord.rdwr(file);
                    }
                   
                    uint8 waiting_time_count = iter.get_current_value().times.get_count();
                    file->rdwr_byte(waiting_time_count);
                    ITERATE(iter.get_current_value().times,i)
                    {
                        // Store each waiting time
                        uint16 current_time = iter.access_current_value().times.get_element(i);
                        file->rdwr_short(current_time);
                    }

                    if(file->get_experimental_version() >= 9)
                    {
                        waiting_time_set wt = iter.get_current_value();
                        file->rdwr_byte(wt.month);
                    }
                }
                halt.set_id(0);
            }

            else
            {
                uint16 halts_count;
                file->rdwr_short(halts_count);
                halthandle_t halt;
                for(uint16 k = 0; k < halts_count; k ++)
                {
                    if(file->get_experimental_version() >= 10)
                    {
                        uint16 id;
                        file->rdwr_short(id);
                        halt.set_id(id);
                    }
                    else
                    {
                        koord halt_position;
                        halt_position.rdwr(file);
                        halt = welt->get_halt_koord_index(halt_position);
                    }   

                    if(halt.is_bound())
                    {
                        fixed_list_tpl<uint16, 16> list;
                        uint8 month;
                        waiting_time_set set;
                        uint8 waiting_time_count;
                        file->rdwr_byte(waiting_time_count);
                        for(uint8 j = 0; j < waiting_time_count; j ++)
                        {
                            uint16 current_time;
                            file->rdwr_short(current_time);
                            list.add_to_tail(current_time);
                        }
                        if(file->get_experimental_version() >= 9)
                        {
                            file->rdwr_byte(month);
                        }
                        else
                        {
                            month = 0;
                        }
                        set.month = month;
                        set.times = list;
                        waiting_times[i].put(halt.get_id(), set);
                    }
                    else
                    {
                        // The list was not properly saved.
                        uint8 waiting_time_count;
                        file->rdwr_byte(waiting_time_count);
                        for(uint8 j = 0; j < waiting_time_count; j ++)
                        {
                            uint16 current_time;
                            file->rdwr_short(current_time);
                        }
                       
                        if(file->get_experimental_version() >= 9)
                        {
                            uint8 month;
                            file->rdwr_byte(month);
                        }
                    }
                }
                halt.set_id(0);
            }
        }

This creates halthandles and assigns IDs to them at loading/saving time. When I comment out the code for if the Experimental version is greater than 10, the game will fail on loading with a quickstone error in the self = halthandle_t(halt_id, this) line complaining that number in halt_id has already been assigned. However, it will not always do this: if the game is loaded paused then saved without being unpaused, it works fine. Crucially, when I do this, I also do not get desyncs. Thoughts on the relevance of this would be appreciated, too.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: Dwachs on February 10, 2012, 11:52:19 AM
Your exceprt regarding saving of waiting time seems to be unrelated to the quickstone errors. If you only access the id's of a halthandle, you only change an index into the pointer table hold by the quickstone. Only the constructors 'self = halthandle(...)' will touch the pointer table, too.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 10, 2012, 12:05:16 PM
So creating a new halthandle object then assigning an ID to it does not affect the pointer table (which is, I assume, where a list of all the IDs is kept such that the program knows if one tries to assign the same ID twice)?
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: Dwachs on February 10, 2012, 12:30:39 PM
It depends on the constructor. This code
Code: [Select]
halt = halthandle_t();
halt.set_id(23908);
would not do any harm to the pointer table and would not contribute to the fatal error you observe. On the other hand, the code
Code: [Select]
halt = halthandle_t(this);
would create a new entry in the pointer table and potentially create such errors.

During reloading of a savegame, all calls to halthandles to register station pointers should be of one of the types
Code: [Select]
halt = halthandle_t(this);
or
Code: [Select]
halt = halthandle_t(this, halt_id);
If you mix both types of calls during one loading operation you will get those errrors sooner or later.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: Junna on February 12, 2012, 11:08:04 AM
May I suggest the server be paused/off while this issue is worked upon to prevent time passing whenever someone actually manages to get on for a short while whereas most cannot? Lindley is likely to be going to go bankrupt otherwise.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: wlindley on February 12, 2012, 02:29:45 PM
Indeed, I have been able to connect for a few seconds now and then, this week.  Enough time to say Arrrgh but not long enough to do anything about it.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 12, 2012, 02:32:37 PM
Ohh dear. I'm sorry about this - I am currently staying with my parents and don't have the login password for the server with me. I shall just have to ask people politely not to log in until I get this fixed, I think.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on February 12, 2012, 03:31:22 PM
don't worry about lindley lines, I fixed that. It now makes positive income.

btw, it seems like it will desync on the first try all the time, but the 2nd and 3rd try, there is a chance that you will be connected for a long time. Also, on the 3rd try, it may crash. It seems almost guaranteed that on the 4th try the game will crash. Also, it does not seem possible for more than one player to play at once.

I was able to log in long enough to add a temporary line alongside lindley's where it was losing money the most, due to a build up of about 15,000 pax.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 13, 2012, 12:32:29 AM
Hmm - the desync bug is not caused by the quickstone bug after all. I have managed to fix the quickstone crashes in my 10.x branch (I think), with the exception of the one in Sdog's game no. 176 (in which the saved game is corrupted - hopefully, the game will no longer generate corrupted saved games, however); but the desyncs persist.

Testing back to version 10.5, loading Sdog's game 162 then connecting to it allows the game to run without desyncs. However, if I then save 162 locally on the server and re-load it again, a client connecting to it will desync on connexion. This is all without producing the quickstone errors that, it seems, entirely coincidentally correlated with the desync errors in my earlier tests. I shall have to continue to look at other possible causes rather painstakingly, but it might take a while. In the meantime, if anyone would like to assist by looking through the code for any possible sync-related errors, I should be most grateful!
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: wlindley on February 13, 2012, 01:39:55 PM
I was able to log in long enough to add a temporary line alongside lindley's where it was losing money the most, due to a build up of about 15,000 pax.

Much obliged, although I believe the cricket fans might rather object to an elevated railway right through the pitch. :o

Also, the latest compile (with the convoy ID update) seems to have solved the desync errors here.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 13, 2012, 08:08:48 PM
Also, the latest compile (with the convoy ID update) seems to have solved the desync errors here.

Hmm, that's odd for two reasons: (1) it didn't solve the desync errors in my tests; and (2) I never pushed a version where this change had any effect, as it is only operative if the Experimental version number is 11 or higher.

Can you elaborate on how you tested this?
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: wlindley on February 14, 2012, 01:00:58 AM
With the version of 5 February, I get a desync after 1 or 2 seconds.  The version compiled this morning stays connected for as long as I have left it running.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on February 14, 2012, 01:11:36 AM
Much obliged, although I believe the cricket fans might rather object to an elevated railway right through the pitch. :o

Also, the latest compile (with the convoy ID update) seems to have solved the desync errors here.
yeah, just tell me when you don't need it anymore. I don't have any interest in expanding, but would rather improve my lines for now.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 14, 2012, 01:26:37 AM
With the version of 5 February, I get a desync after 1 or 2 seconds.  The version compiled this morning stays connected for as long as I have left it running.

Compiled from the 10.x branch? You are able to connect to the server with that...? They're not the same version. Hmm - very odd.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 18, 2012, 11:41:39 PM
I have made significant progress in tracking down the desync issue. Loading Sdog's game "173", setting a breakpoint at line no. 1167 in simhalt.cc (in the current 10.x code) with the condition: "TEST_id == 1039 && TEST_id_dest == 2018" reveals that, seemingly at random, sometimes on loading the game and re-calculating the routes with "reroute_goods()", passengers at Brighton Stop bound for Mansfield Small Cricket Ground Stop will route via Brighton Town Hall Stop, with a journey time ("TEST_ROUTE") of 4496 (i.e., 449.6 minutes, or 7:24h), but sometimes the passengers will route instead via Brighton Railway Station, with a journey time of 4175 (417.5 minutes, or 6:58h). When run as a client/server pair, the server will usually retain the longer route, whereas the client will usually re-route to the shorter route. The first leg of the shorter route is achieved on foot using the walking connexions setting, so pedestrians will be generated (using the random number generator) during the recalculating routes phase of loading, and the random seeds will be desynchronised before even the first step.

It is not currently clear why this divergence is occurring. Passengers should always take the fastest route. Checking those particular stations on the map, both routes appear valid: one involves taking a 'bus a short distance to a very nearby stop, and then taking a longer 'bus journey to a dock, taking a ship accross a large lake, and then another 'bus to the final destination; the other involves walking to a local station, taking a slow train to the same dock, and the same route as before thereafter.

The difficulty with tracking down the issue is that it only appears on larger maps, yet it is very difficult to isolate particular instances of anything in the debugger on a large map when there are so many different things. Any coding assistance in tracking this down would be most welcome.

Edit: Further testing has confirmed that the walking connexions feature is not itself to blame for this issue: disabling the feature entirely still results in near instant desyncs on the 173 map.

Edit 2: A copy of Sdog's game no. 173 is available here (http://simutrans-germany.com/files/upload/173.sve) for testing.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: sdog on February 19, 2012, 02:36:28 AM
in fact i had some very strange routing of pax occuring. About 30k pax in Shrewsbury Station rerouted frequently to completely different via destinations seemingly random. It included Sheffield (1000 tiles to the east), Taunton (500+ tiles to the NW), a stop within Shrewsbury, Shrewsbury Abbey station and sometimes distributed over a very large number of destinations none exceeding a few thousand pax) They fluctuated within a couple of game seconds (this is every time i unpaused and re-paused again i saw a different destination).

James please bear in mind that likely the total travel time is determined by the wait time at completely overcrowded stations. If this changes eg. by a convoy departing somewhere it can instantly change the wait times. The map has a very densely interconnected network leading to a complex system. (see how it is initially described by simple equations?) My guess is we see chaos here!
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 19, 2012, 11:37:39 AM
Ahh, but my tests should give deterministic results, because exactly the same saved game is being loaded on each occasion, and the calculation of paths is always performed immediately after loading. However complicated the network, therefore, the results should be identical every time.

Edit: Further testing appears to show that waiting times might be the issue, although it is not entirely clear, as I cannot immediately link the discrepancy to this particular route.

At Bristol station, the only route to Bristol East Stop is by line no. 208, a 'bus route. Loading the game, the waiting time for Bristol East Stop is sometimes unknown (which defaults to about 2 minutes) or 17 minutes and 18 seconds. This discrepancy correlates exactly with the discrepancy with passenger routing from Brighton Stop to Mansfield Small Cricket Ground Stop: when passengers walk to the station and take the train, the waiting time between Bristol and Bristol East is unknown, whereas when they take the 'bus to the Town Hall stop and change for another 'bus to the dock, the waiting time is 17:18.

Edit: Additionally, it seems, saving and loading has a predictable effect on this. Starting the game afresh (with the demo map automatically loaded), then loading 173.sve, the waiting time from Bristol Station to Bristol East will always be unknown. Re-loading 173.sve without quitting the game and starting again will always give the 17:18 waiting time for Bristol East.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: Junna on February 19, 2012, 03:29:50 PM
Edit: Additionally, it seems, saving and loading has a predictable effect on this. Starting the game afresh (with the demo map automatically loaded), then loading 173.sve, the waiting time from Bristol Station to Bristol East will always be unknown. Re-loading 173.sve without quitting the game and starting again will always give the 17:18 waiting time for Bristol East.

This would explain why it always desyncs the first time you join the online game and why you can manage to stay on when you re-load from there.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 19, 2012, 05:16:01 PM
I think that I have - eventually - solved this: see the 10.x branch. The problem was with waiting times not loading correctly after being saved.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on February 19, 2012, 05:48:13 PM
that's most excellent news. :)
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 19, 2012, 05:52:31 PM
Yes, you will be able to return to world domination once I have deployed this latest version! ;-)
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on February 19, 2012, 06:25:02 PM
looking at the finances of everyone, network scale and how crowded it is around the cities, I don't think it's possible for any one of the major three to out compete any of the others. Unless, maybe, there is cooperation between two of the three, but even then, I doubt it.

We all just earn so much interest, that we would have to suffer substantial loses for there to be any hope of bankruptcy.
by substantial, I mean something along the lines of $50 million yearly, at least.


Oh, for anyone starting up a new company. There are still plenty of cities and towns that need connecting, and it is possible to make a profit even with only $250,000 in starting money. There are quite a few clusters of cities in the north east and south east corners that can easily result in profits. The LMS push-pull sets are more than adequate for this and the diesel buses are not bad either.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: Milko on February 19, 2012, 08:51:26 PM
Hello

I think that I have - eventually - solved this: see the 10.x branch. The problem was with waiting times not loading correctly after being saved.

James, the killer bug!  :) Great!

Giuseppe
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 19, 2012, 09:51:49 PM
The server has now been restarted with 10.10, and no longer seems to desync. Thank you everyone for your patience!
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: rsdworker on February 19, 2012, 10:24:45 PM
great i will come back to playing with with players :D
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 19, 2012, 10:31:11 PM
Excellent!
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: sdog on February 19, 2012, 11:21:56 PM
what's the passenger level at the server game. (still afraid to even look at it, online play is too addictive to me)

I notice that even at my game with the old low passenger level there aren't villages small enough to have use for any of the low capacity trains, like that 1 unit dmu (class 101? can't check here right now).
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 19, 2012, 11:36:19 PM
It is currently 12, although it will be reduced to 11 for the next release of the pakset, as well as reducing the journey time tolerance (which will reduce passenger numbers further).
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: sdog on February 19, 2012, 11:46:40 PM
Oh yes, journey times don't play a role at all anymore. In late 2009 i think you had the best balance. It was a very narrow edge for profitability, local transport had to be very well optimised for low travel times, else massive refund could cause a profitable company to quickly drop to inprofitability. (i managed only with three in about 20 tries to succeed for more than a decade, and only one for a few decades.)


Furthermore saturation of lines is a much bigger problem than profitability. Car usership didn't change a lot, as population growth over-compensated it.


(sorry, this is not related to bridgewater brunel, but is related to the topic)
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 19, 2012, 11:47:57 PM
Oh yes, journey times don't play a role at all anymore. In late 2009 i think you had the best balance. It was a very narrow edge for profitability, local transport had to be very well optimised for low travel times, else massive refund could cause a profitable company to quickly drop to inprofitability. (i managed only with three in about 20 tries to succeed for more than a decade, and only one for a few decades.)

Hmm - this wasn't really quite right, however, as there was little use for high capacity vehicles. What is needed is a setup where low, medium and high capacity transport networks are all useful in their place.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: sdog on February 19, 2012, 11:59:13 PM
That old state is perhaps a bit too extreme. But don't forget we have clustering of cities now and diferent travel distances of pax. The high capacity long range stock will come to play to connect distant population centres. If i remember correctly the savegames in the 17x range were saved with pax factor 8 or 10. (can't check right now). Pax per tile within the allowable weight is the only thing that counts for me when buying convoys.

Going to larger large cities in the 30k to 40k range provides a very high demand on the main line. Having smaller large cities with a higher pax factor would also work, but then the little villages provide considerably more pax. (and mail, there's no mail in that game, as it was completely impossible to move that sheer amount, without building a completely new rail network)
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: rsdworker on February 20, 2012, 12:14:25 AM
ok joining game now :D
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: dustNbone on February 20, 2012, 12:29:51 AM
I don't find that mainline saturation is a problem yet (1910ish).  I'm still only using 5 tile trains and generally have a fair amount of capacity left, though traffic is quite high on some lines they're still not getting in eachothers way thus affecting average speed.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 20, 2012, 01:23:48 AM
Bear in mind that your main lines should not be anywhere near saturated with passenger traffic: there should be slow freight traffic interspersed with the fast passenger traffic, as there was in reality, and some lines would, in reality, have far lower utilisation than others (and therefore have lower profitability).

Urban short distance railways are another matter, and they might very well be saturated with passengers.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on February 20, 2012, 01:43:23 AM
I'm sure in real life, there would have been competition by increasing service quality or decreasing fairs.

PS. I will need road works near elford and nightworth
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: AP on February 20, 2012, 08:03:58 AM
Oh, for anyone starting up a new company. There are still plenty of cities and towns that need connecting, and it is possible to make a profit even with only $250,000 in starting money. There are quite a few clusters of cities in the north east and south east corners that can easily result in profits. The LMS push-pull sets are more than adequate for this and the diesel buses are not bad either.
I'm still practicing in my sandbox, trying to get a feel for what makes experimental 'tick'...
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on February 20, 2012, 08:57:39 AM
I have setup a pair of starter companies that make a profit and can easily be expanded.

Hardden & Colbrook Railways is still small, but offers plenty of hope. It doesn't earn much, but is quite stable with a sizable budget for expansion.

Salford Railways only runs one line, but has found its niche in a rather busy area. It earns quite a lot, thanks to its fast service through a crowded corridor, but the competition is fierce and its foundations are rather shaky at the moment.

For both companies, password is: password



oh, and james, I think it would be best if you can unlock player 1, 2, 3.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 20, 2012, 09:38:50 PM
PS. I will need road works near elford and nightworth

Elford I have done, but I am not sure what you want at Nightworth...?
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: rsdworker on February 21, 2012, 12:34:54 PM
I have setup a pair of starter companies that make a profit and can easily be expanded.

Hardden & Colbrook Railways is still small, but offers plenty of hope. It doesn't earn much, but is quite stable with a sizable budget for expansion.

Salford Railways only runs one line, but has found its niche in a rather busy area. It earns quite a lot, thanks to its fast service through a crowded corridor, but the competition is fierce and its foundations are rather shaky at the moment.

For both companies, password is: password



oh, and james, I think it would be best if you can unlock player 1, 2, 3.

my former company - city tramway which is no longer operates because i was abandoned company because desyncs - so its seems working pefectly

edit 1: city tramway is up for takeover - password is password
so i don't see those two companies - salford also H+C on player list

edit 2:
scratch that - i taken over H+C company :)

edit 3:
hmm - why password is not working - there no message to confirm the password is correct or wrong

edit: 4: - ahh works now - but desync is small but better now

edit 5: - jamespotts - you need sort the stuck trams messages - its annoying

btw - i extended the line to Turnthorne
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: Carl on February 21, 2012, 03:55:54 PM
Just checking this out for the first time today -- I grabbed one of the empty slots. Trying to connect some of the remaining towns in the East.
(My first time on pak128britain.ex too, I should say, so there's a good chance that my business will not be a success! :) )

A couple of desyncs so far, but I'm willing to bet that's due to problems at my end.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: wlindley on February 21, 2012, 04:57:53 PM
is it me, or does doing anything (laying even one tile of track, etc.) take two or three minutes?  i do not get desyncs now, but the game is unplayable.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: rsdworker on February 21, 2012, 05:08:02 PM
is it me, or does doing anything (laying even one tile of track, etc.) take two or three minutes?  i do not get desyncs now, but the game is unplayable.
yes i did play earlier - i experinced simllar problem
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: Carl on February 21, 2012, 05:09:50 PM
I find it to be more like 2 seconds than 2 minutes, but there's certainly a delay.

---

Edit:
Dear me, looks like I completely misunderstood rail weight limits, meaning I bought far too expensive track. Without that, I'd be making a profit... :(


---


Edit 2:
I've aborted Bakerlink, chalking it all up to experience...
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on February 21, 2012, 06:15:25 PM
I get around 2~4 seconds of delay.

I wish I could unlock player slots 1 and 3...

I mean, I could have sworn I started up 3 and put "password" as the password, but alas, that's not it.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: AP on February 21, 2012, 08:09:40 PM
Dear me, looks like I completely misunderstood rail weight limits, meaning I bought far too expensive track. Without that, I'd be making a profit... :(
Weight limits can be a bit of a pain, frankly. I had the reverse trouble with my second attempt on the map- the empty train could run over the tracks, but the weight of the cargo when it was loaded was too much for the track to carry.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: dustNbone on February 21, 2012, 08:19:34 PM
Yeah the double decker buses do similar things.  As well as being a bit underpowered, they can't get up a hill with more than 15 pax or so.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: AP on February 21, 2012, 09:16:09 PM
Yeah the double decker buses do similar things.  As well as being a bit underpowered, they can't get up a hill with more than 15 pax or so.
James has really got the realism down to a tee!  ;D
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 21, 2012, 11:17:23 PM
Yeah the double decker buses do similar things.  As well as being a bit underpowered, they can't get up a hill with more than 15 pax or so.

May I ask which specific 'buses you were having trouble with? If the petrol electric, they were underpowered in reality...

Edit: As to the delay, I had increased server_frames_ahead to 6 when we had the earlier desyncs before I realised that the problem was due to a bug. I have re-set it to 4 now, the default; let me know if (after the next save/reload) the problem gets any better.

Edit 2: I have unlocked two dud players, "Player 1" and "Player 3", so there are now two free slots if anyone would like to join in and give the existing players a run for their money...
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on February 22, 2012, 01:19:21 AM
we have a lot of money, btw ;)
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 22, 2012, 01:26:39 AM
I noticed! Rebalancing is required, I think!
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: Junna on February 22, 2012, 10:04:49 PM
It no longer seems to appear in the announce and the in-game server list?
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 22, 2012, 10:13:16 PM
It does appear on the listings website (http://servers.experimental.simutrans.org/list). There is an issue with the exact name of the pakset (I am not sure why it has arisen now and not before) meaning that it does not currently appear on the in-game list unless one checks "show all" for Windows clients - this should be resolved for the next release of Pak128.Britain-Ex.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on February 27, 2012, 03:58:47 AM
in 10.9 'show all' was selected by default, but in 10.10, it's not. I think that's why it doesn't show up.


the server seems to have crashed. I got a protocol error.

I would guess that the finances ran out of digits at $10,800,000,000.00 or something.

we were starting to get negative interest at that range.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: dustNbone on February 27, 2012, 04:11:38 AM
Yeah there's been a bit of weirdness.  In and around the time I stopped gaining money (I thought I was being supertaxed :)  the game decided to seemingly at random delete various stops from my schedules.  It seemed to mainly be schedules that I'd most recently created or modified, some just had a stop or two removed and some actually had all their stops removed, sending all the vehicles to depot.  This obviously created a very large mess which I have only partly repaired thus far.  I'm thinking we broke something.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 27, 2012, 09:58:27 AM
Hmm, that is extremely odd. The server still seems to be up according to the listings server, but I shall take a note about what appears to be an integer overflow in the interest calculations. Sorry about that difficulty!

I'm not quite sure how to track down the issue with deleting halts from schedules - did this only happen once? It is almost impossible to find a bug without being able to reproduce it reliably.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: dustNbone on February 27, 2012, 10:07:30 AM
Yeah it seemed to just happen in one lump.  Doesn't seem to be happening again.  It was very odd behavior :)
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 27, 2012, 10:51:02 AM
Very curious indeed... that almost sounds like memory corruption, although a whole lot of those sorts of bugs had - I thought - been remedied of late. But there are whole new sorts of iterators in the latest Standard nightlies that I am painstakingly merging into the the 111.1-merge Experimental branch on Github (and the time that it is taking to do that is why I have not yet had a chance to update the -devel branch with the excellent looking patches from Carl and Richard), which I think were intended partly to address possible memory corruption issues, so it might be that this problem magically disappears in the next major version - we shall have to see.

(The interest rate bug is a different matter and will need its own fix).
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on February 28, 2012, 03:17:48 AM
well, the server is working now.

the passenger network itself seems to be nearly overloaded. I think it would be completely overloaded if motorization wasn't in full effect.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: dustNbone on February 28, 2012, 03:42:40 AM
Lol yeah I had a few line go to depot and stop working.  Resuming services now :)   The southern undersea tunnel wasn't running so that would make alot of traffic elsewhere.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: Carl on February 28, 2012, 08:25:54 AM
I've been absent since the 1920s, and the towns I serve have completely outgrown my rudimentary network. One really has to keep a regular eye on this!
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: Milko on February 28, 2012, 09:28:27 AM
Hello

Pak128Brit has a lack of medium-density buildings in the era 1920 - 1940.

http://forum.simutrans.com/index.php?topic=7721.msg73307#msg73307 (http://forum.simutrans.com/index.php?topic=7721.msg73307#msg73307)

Giuseppe
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: wlindley on February 28, 2012, 03:02:50 PM
With several of the versions (from 5 Feb, and 8 Feb) that formerly worked, I now get to "Calculating Paths," the bar goes all the way across, and... nothing.  CPU at 0%.  Just sits there.  What with lag and disconnect problems, I have not been able to play since the 1880s.  What now?
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: dustNbone on February 28, 2012, 04:13:40 PM
I offer $10 Billion simutrans dollars to the first person with a good answer for wlindleys question.  Since I'm not allowed to have more than $10.8 Billion :)
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: Milko on February 28, 2012, 04:31:15 PM
Hello wlindley

With several of the versions (from 5 Feb, and 8 Feb) that formerly worked, I now get to "Calculating Paths," the bar goes all the way across, and... nothing.  CPU at 0%.  Just sits there.  What with lag and disconnect problems, I have not been able to play since the 1880s.  What now?

Are you connecting to the server using a self compiled build (different from 10.10 release version)?

Giuseppe
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: wlindley on February 28, 2012, 05:15:57 PM
Yes, I am -- the announcement post (http://forum.simutrans.com/index.php?topic=9320.0) still does not have 64-bit Linux binaries.  Or can I find that somewhere?
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on February 28, 2012, 05:33:48 PM
there seems to be more crashes in 1959. The only thing different, that I've noticed, is the new power plant.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: dustNbone on February 28, 2012, 05:58:26 PM
Yes, I am -- the announcement post (http://forum.simutrans.com/index.php?topic=9320.0) still does not have 64-bit Linux binaries.  Or can I find that somewhere?

Here's a link to the one I built (and use)
http://dl.dropbox.com/u/38204997/simutrans/stexp1010
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 28, 2012, 10:56:20 PM
I am very sorry about the Linux problems - for some reason, the automatic Linux builds refuse to compile for version 10.10, although I can compile the server build without difficulty on my server. If anyone has any idea what the problem might be, I should be most grateful. I can post the error messages if that would help.

DustNbone - would you mind if I linked to your version on the official release thread?

AEO - can you elaborate on the nature of these crashes - are they preceded by the odd behaviour suggestive of memory corruption discussed above?
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on February 28, 2012, 11:49:58 PM
the crashes seem to come most at the beginning of the month.
that particular one in 1959 caused a lot of convoys to get stuck.

I've also noticed that one ship, when replaced, was sitting in the depot with something like 20 copies of its line in the schedule list right before the crash. It was like "stop A, stop B, stop A, Stop A...".
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 29, 2012, 12:03:33 AM
Hmm. Looks like memory corruption in collection classes again.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: dustNbone on February 29, 2012, 02:52:06 AM
Yeah I don't mind, should I keep it on dropbox or will you put a copy somewhere else and link to that?
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on February 29, 2012, 07:26:33 AM
server seems to lag or even crash when replacing electrification en masse
same deal with removing tram tracks en masse.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: Milko on February 29, 2012, 09:13:50 AM


Yes, I am -- the announcement post (http://forum.simutrans.com/index.php?topic=9320.0) still does not have 64-bit Linux binaries.  Or can I find that somewhere?


The server and the client versions must be the same. Using different versions increases the disconnection because the code that runs on the server is different from the code that runs on the client and the results of the calculations are so different. I think you have to build the repository "master" from github to use online games...

Could this be the reason?

Giuseppe
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: wlindley on February 29, 2012, 02:43:58 PM
I did a 'git checkout master' 'git pull' and 'make -j3' ... with that binary I can now connect, wait about two minutes for initialization, and play for a couple seconds before everything freezes.  No errors, just a freeze.   At least I have been able to replace some of my circa 1870s  horse-and-carriage with a bus, but at this pace I will never be able to upgrade the rail lines.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 29, 2012, 09:29:41 PM
dustNbone - that is very kind. I will take a copy, if you don't mind.

AEO - noted: I'm not sure what might be causing that.

Wlindley - may I ask what the significance of "-j3" is? The server version is built without this switch.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: wlindley on February 29, 2012, 10:15:02 PM
j3 just means to compile up to three modules at once.  Dependencies are still respected.  The same thing happens without that switch.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on February 29, 2012, 10:20:33 PM
Interesting. What version of GCC are you using, and from which branch are you pulling?
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: Junna on March 01, 2012, 10:23:59 AM
Desyncs and crashes are near constant making it frustratingly unplayable. I bet it's the city cars.

Changing electrification also seems to contribute.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: dustNbone on March 01, 2012, 10:43:01 AM
Yeah I've noticed even in single player the city cars seem to just brutalize the CPU starting around the mid 60s.  Is it simulating their trips or just animating them?
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on March 01, 2012, 10:47:38 AM
Hmm, I have also noticed issues with desyncs (at the beginning of the month mostly, but also at other times) and some crashes. Not entirely sure why at present, as have not had time to investigate. A problem is that there are known memory corruption issues which should be fixed when I am able to merge all the latest changes from Standard 111.2.1, on which I have been working. Until I do that, it will be hard to tell which issues are and which issues are not related to this problem.

As to the private cars, their trips are simulated, but in a rather abstract way, and the private car graphics are not themselves simulating car trips, but just wandering around more or less aimlessly in numbers directly proportionate to the number of actual private car trips.

Edit: Incidentally, if anyone would like to help to track down this issue, I should be very grateful if anyone could run the current server game with profiling enabled to see which parts of the code are currently the most CPU intensive.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on March 04, 2012, 12:52:12 AM
server seems to have crashed from
Code: [Select]
minivec_tpl<T>::resize()
new size 256 too large
(>255).
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on March 10, 2012, 02:01:20 AM
Apologies for not having got to this before - I had been preoccupied with trying to deal with the pesky iterator issues in the merge version. I have now found and fixed the cause of this crash in the 10.x branch - would anyone care to test whether this fix causes unintended anomalies?

Thank you everyone for your patience!
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: dustNbone on March 10, 2012, 04:55:57 AM
I built and ran the latest save I had locally from the server game, October 1977 game time and ran it for about 2 years and it seems fine.  Saved and reloaded a few times, it still won't let us have more than 10.7 billion or so but that's enough money I guess.  It would be nice if it would at least rollover to 0 maybe at that point (maybe as an option).  Maybe not realistic but at least you can keep progressing.

Oh and here http://dl.dropbox.com/u/38204997/simutrans/stexp10xmar9x64 is a link to my x64 Linux build if anyone needs it.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on March 10, 2012, 01:34:05 PM
The 10.7 billion issue, I suspect, is indirectly caused by the negative interest bug, which I have now fixed in 10.x. It's rather difficult to test, as it requires the game to run for a rather long time in the right conditions, however.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on March 25, 2012, 04:46:03 PM
Apologies for the recent downtime - this is now running again with the latest 10.11 release. In view of the problems discussed here (http://forum.simutrans.com/index.php?topic=9595.0), I have enabled assume_everywhere_connected_by_road, so there will be much more private car competition to and from all cities than there was hitherto (when the road connexion databases were, due to a bug, no updated after they were first created).

Thank you all for your patience, and happy playing!
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: rsdworker on March 25, 2012, 05:41:47 PM
i tried to get on but problem is simtrans hangs up when loading the server - its seems something wrong?

i updated the simtrans exp to 10.11
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on March 25, 2012, 07:04:58 PM
Hmm - I can't reproduce this. It does take a fair while to download the game, but it is rather large.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: rsdworker on March 25, 2012, 07:32:01 PM
Hmm - I can't reproduce this. It does take a fair while to download the game, but it is rather large.
before that update - its didn't take long time to load world
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on March 25, 2012, 07:32:53 PM
Hmm. It does seem to take a little longer, although it's not clear why at present. However, it does eventually load.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: rsdworker on March 25, 2012, 08:08:10 PM
great - took around 5 minutes to log in the server
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on March 25, 2012, 09:28:31 PM
I was dropped the first time I connected, but it seems to work now.
I think the general slowness is caused by the lack of processor power on the server side, because if I load the game as a local copy on my computer, it becomes very unresponsive as well.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: rsdworker on March 25, 2012, 09:46:00 PM
I was dropped the first time I connected, but it seems to work now.
I think the general slowness is caused by the lack of processor power on the server side, because if I load the game as a local copy on my computer, it becomes very unresponsive as well.
- that's what i thought when i tried to login to server - its was very slow well
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on March 25, 2012, 09:48:34 PM
private cars are spawning where they shouldn't.

for instance: PEX Catingfield station
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: rsdworker on March 25, 2012, 09:50:43 PM
private cars are spawning where they shouldn't.

for instance: PEX Catingfield station
that's true - i saw alot of traffic and its was very slow
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on March 25, 2012, 09:54:53 PM
AEO - can you repost that as a bug report? It's not really specific to the server. Thank you for reporting it, though.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on March 26, 2012, 10:21:03 AM
There seems to be some instability upon quitting the game. Sometimes it reverts to an older save and other times it works fine.

public cars are piling up in the diagonals and this sometimes causes grade crossings to jam. some designs are worse than others. Addham seems to be particularly bad.

The input delay on the client end might be caused by the client lagging behind the server. For me, this is a gradual build up. I've noticed that when logging on with a second computer, both underpowered, there was always a delay for the computer that's been logged on. The computer that just logged on will be paused with the server, waiting for the other client to catch up.



Electricity works great now.It just needs to be connected to cities :D
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: Vonjo on March 27, 2012, 05:54:31 PM
I think 18 MB to download this network game is really too big.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on March 27, 2012, 07:22:04 PM
once the game loads, you have to wait for your computer to catchup to the server. How fast that happens seems to be related to how fast your computer is.

You can give a lot of command inputs while the game is paused and they'll either all go through, or you'll be desynced.

Waiting times for sync are quite long.
Unloading the map or quitting the game after the server game is loaded also takes a ton of time.

For my computer, which is Core 2 duo 1.86ghz and 2GB ram, the waiting time is anywhere between 20 and 40mins.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on March 27, 2012, 10:52:17 PM
Hmm, that's rather a long time. The difference in performance seems to have occurred after the update to 10.11. The only thing that I can think of is that it is the number of private cars, which would have increased as a result of setting assume_everywhere_connected_by_road=1. I shall have to perform some tests if I have time to see whether this is indeed the problem, and shall have to consider, if it is, what to do about it, and, potentially, how to reduce the amount of memory taken by each private car.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on March 30, 2012, 01:36:25 PM
I've checked the config.tab, but it doesn't seem like assume_everywhere_connected_by_road=1 is enabled. It's still flagged to 0.

I've tested the performance difference between the two as a local copy. Game speed, at best, is 0.31 before it becomes unresponsive for both.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on April 04, 2012, 11:22:25 PM
I think it might be best to restart the map with more conservative settings, as it is unplayable right now.

Say, with a smaller area and around 450 cities, 300 industries and 500 attractions?

I've added some maps to the list of good or interesting maps that should result in a similar game play experience to the current map.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on April 04, 2012, 11:29:49 PM
Sorry for not having replied to this before. The performance issue here is not a factor of the private car routes feature, as that is disabled (even if not in the simuconf.tab file, it is indeed disabled in the .sve file, as will be seen if you download the game, open it in single player mode, and press the "i" key: "assume_everywhere_connected_by_road" will be checked).

The problem appears to be specific to the new version, with the same saved game performing much more slowly in 10.11 than in 10.10. This looks as though it is probably a bug, which I shall have to look into when I clear the merge issues induced backlog that my work schedule this week is not making go away any faster.

Thank you for reporting the issues, though: that is helpful.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on April 07, 2012, 04:25:48 PM
here is a slightly touched up #9763. Main difference from stock are the connections of seas and lakes. Map has been rotated as well.
http://dl.dropbox.com/u/17111233/Trial%2010.11.rar (http://dl.dropbox.com/u/17111233/Trial%2010.11.rar)

It should be decently challenging to play if used. It also skips ahead to 1835 instead of 1825, because there really is not much that can be used until 1835.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on April 07, 2012, 06:03:26 PM
Hmm - I'd rather be starting at 1800, actually, to make full use of canals.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on April 08, 2012, 12:55:53 AM
redone in 1800.
Trees also removed, which decreased file size to about 1/5

http://dl.dropbox.com/u/17111233/Trial%20v2%2010.11.rar (http://dl.dropbox.com/u/17111233/Trial%20v2%2010.11.rar)

I edited the config file to use macadam road to save hassle down the line.


---
I think, as long as no one adds too many cities, it should remain playable until the end.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on April 15, 2012, 07:16:11 PM
I have restarted the server with a slightly modified version of the current game, removing private cars, which were causing performance problems (see here (http://forum.simutrans.com/index.php?topic=9720) for discussion). The performance of the server is now drastically improved, and readily playable. Apologies for the difficulties that have obtained in the meantime.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: rsdworker on April 15, 2012, 07:23:06 PM
great its will be good to try - i don't understand why private cars causing problems - its needs looked at it
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on April 15, 2012, 07:31:46 PM
Indeed - I am in the process of looking at it! To clarify, private cars will still compete with player transport for passengers, and will still cause congestion to register on the graphs in the cities, but the private car graphics will not be displayed in the game.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on April 17, 2012, 11:30:15 PM
server works fantastically well without the private cars.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on April 17, 2012, 11:44:29 PM
Splendid! Just looking now into the performance issue with private cars.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: DatDamnDom on April 19, 2012, 12:49:55 AM
*Noob alert*

Hello folks.  Recently rediscovered the game.  When I last put it down I think the latest stable was in the mid 80s, or something.

I like the improvements, and I like the idea of online play.  But the online interface involves elements that are unfamiliar to me.

Could someone help me navigate what's involved in logging into an online game?  I see the server lists, but I have no idea where the "url" is (I know my way around computers but I'm no software engineer) and the dialogue box where I type in server info manually isn't very illuminating in its description.

I hope I placed this comment in the correct thread.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on April 19, 2012, 01:47:39 AM
Hello! Welcome to the forums, and welcome back to Simutrans.

There is presently a small bug which prevents the server game from being shown in the list unless you check the "Show all" box next to the "add server" button. Click "Show all", then click the blank text box above it, and hopefully you will see the URL of the server. Once you have selected that, providing that you have the correct version of the game and the pakset, the "Play online" box will become available. Click that, and the server game will load. Note that, because the map is very big, it will take a good few seconds to download. Once the map has loaded, you will need to select your company (click the disk icon in the top left, select "players", and find an empty slot), then click on the green button to rename your company and give it a password so that only you can control it. Note that you can change the company name and password at any time. Once you have done that, you are ready to start playing.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: DatDamnDom on April 19, 2012, 02:31:47 AM
Thank you very kindly for your timely reply.

I know what paksets are, but I'm only vaguely familiar with how to change or download them.  (I do note however that from what I can tell, this process is VASTLY streamlined from what I recall.  I've already downloaded and used a 64, a stable 128, and a nightly 128.  I like how they're all concurrently installed, and I need only select which one I want to use at startup.  I remember having to have multiple config files in temp folders and swapping them oout before switching paksets, and using the wrong one would cause my computer to crash.)

I suppose I'll fiddle with it to see if I can discover some of my own answers, but your help here has set me in the right direction it seems.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: dustNbone on April 19, 2012, 04:12:44 AM
Just extract each pakset into its own folder inside the Simutrans folder, the game will search these out and offer a menu at startupi from which you can select from your installed paksets.  Alternately you can specify one from the command line with the -objects switch.  Hope you are successful :)
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: DatDamnDom on April 19, 2012, 04:50:30 AM
This I have actually discovered since my last post, so thank you.   ;D That said, I'm still not keen on figuring out which one is "better," but I anticipate this will come to me with time and practice.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on April 19, 2012, 06:13:12 AM
I've managed to produce a rare bug, according to me, anyways. I've only gotten it once before. Upon replacement, sometimes a convoy will bug out and tell you "depot cannot be found, convoy must be sent there manually", even though it has already been replaced. It will constantly give out this warning as long as the convoy is in service and the error message will only go away when the bugged convoy is withdrawn.I know it's one of the layland DAF 85 arctic mail trucks, but I don't know which one exactly.If you switch to my company, you will see it.


---

Okay, it seems that it was actually the narrow gauge train that had no depot.

---

There seems to be another bug. I replaced class 43 diesels with 365 dual modes. They run over another player's tracks with different electrification and it seems like the route finder can't route the trains over someone else's tracks with differing electrification type. It happens just south of spinford.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: rsdworker on April 20, 2012, 11:08:29 PM
ok i am back on online game - so which company i need take over because i noticed some still have steam trains but i can't find my company who i took over on server

thanks
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on April 20, 2012, 11:35:00 PM
Hmm - what do you mean when you say that you can't find a company?

AEO - will look into that bug when I get the time? Would you mind posting bug reports in their own threads, though? Thank you!
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: rsdworker on April 20, 2012, 11:35:53 PM
Hmm - what do you mean when you say that you can't find a company?

AEO - will look into that bug when I get the time? Would you mind posting bug reports in their own threads, though? Thank you!

i can't remember name of company that i took over but look in previous posts
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on April 20, 2012, 11:45:21 PM
If you don't know which company that you took over, how do you know that it has steam locomotives? Did you know that you can click on a way and see the name of the company that owns it?
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: rsdworker on April 21, 2012, 01:38:03 AM
If you don't know which company that you took over, how do you know that it has steam locomotives? Did you know that you can click on a way and see the name of the company that owns it?
Hardden & Colbrook Railways that company which i took over but i left because server had problems but its not there on player list anymore
and i will find which company has steam trains and try to take over
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on April 21, 2012, 05:42:46 AM
AEO - will look into that bug when I get the time? Would you mind posting bug reports in their own threads, though? Thank you!

It seems to be hard to reproduce on a small local game. I'm starting to think it's related to the stalls when replacing electrification on a big map. Route finder retains the limits of the old convoy, so it assumes the replaced convoy is also diesel.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on April 21, 2012, 11:49:12 AM
Hardden & Colbrook Railways that company which i took over but i left because server had problems but its not there on player list anymore
and i will find which company has steam trains and try to take over

If it's not on the company list any more, I'm afraid that that means that it's been liquidated.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on April 21, 2012, 05:11:59 PM
I would say that it's actually quite difficult to run a profitable passenger or mail line, starting from around 1970. It's really important that your lines are augmented with bus or tram lines. A rail network that only covers the city hall and nearby attractions will quickly run into red.

Currently, I'm renovating the entire western coast, but it's time consuming and tiring to my clicking finger.
At least I now have a fairly good design for running mail and pax lines efficiently.

Pax: grid bus network
Mail: hub and spoke truck network

Works way better than tube trains, for some reason.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on April 21, 2012, 07:58:32 PM
Interesting! When you say that it works better, what do you mean - in what way does it work better?
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on April 22, 2012, 05:54:19 AM
I'm finding that most of the tube lines simply have way too much capacity for the amount of usage they are getting. They were pretty good for about 50 years, but motorization seems to have taken its toll on the tubes as well. Buses with dedicated roads can do the same job better, because they are faster to accelerate and they have a smaller capacity that better matches the demands. What's more, running bus lines also allows room for mail trucks to be run adjacently.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on April 22, 2012, 11:56:13 AM
AEO,

that's a useful insight. I wonder whether this has to do with congestion, and the way in which it is calculated. Congestion is currently based on a "congestion density factor", the size of the town and the number of cars. This results in some very large cities having much less congestion than some smaller cities (Scarmead, for example, has about 1,900 congestion, whereas Workmere has only congestion of 54). The congestion impacts on people's willingness to take public transport. (There is also an insufficient local bias, I think, although that will be rectified in the next version). A better means of calculating congestion might well help to encourage people in larger cities in particular to take public transport, while leaving those in smaller towns to use their car as happily as they do now.

The ultimate solution for congestion calculation would be to calculate actual routes for private cars, have them travel along those routes, and then get actual congestion data from each individual tile (i.e., how long that road vehicles on them have been stationary in the last month), but this would require much work, and I do not know whether it is feasible at reasonable performance given the computational effort that would be involved.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on April 22, 2012, 01:06:20 PM
Hmm - we seem to have a rather late Beeching axe on the West Coast! Were profits too low?
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on April 22, 2012, 01:49:52 PM
oh, it's not a beeching axe.
Profits were low, but it's a full tear down and renovation for the grid-bus line and hub-spoke mail lines.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on April 22, 2012, 02:09:29 PM
Hmm - but the whole coatford area now has no public transport at all.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: dustNbone on April 22, 2012, 04:35:46 PM
Wow, that is one mega project you have undertaken there.  I've noticed that it seems to have attracted a fair bit of traffic to the whole network, keeping me at least near profitability. 

Hoping to get some time to sort out my mess and try to provide some local services in the east.  You're correct as far as passenger transport being minimally profitable after the 1980s, at least on a small enough scale for a new player to succeed.  I tried setting up a new company just to run some local/rural bus routes and it failed miserably (Translink).

The local tube trains are pretty much a hopeless cause as AEO pointed out above, just can't get enough traffic on them, and their performance really seems lackluster on short runs within a city.  Overall, it's worth running local services just to feed the much more profitable long distance high speed ones, but they will run at a considerable loss as far as I can tell.


Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on April 22, 2012, 07:13:13 PM
dust, if you can, could you please remove the bridge that crosses water at feathervale?

james, could you please remove the road at axingworth millborough station?
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: dustNbone on April 23, 2012, 11:49:57 AM
I think I have done it :) If you have any mutually benficial suggestions for where I should start I would be grateful.  My experience with Simutrans is fairly limited and you seem to have a much firmer grasp on these thing from me.  I'm going to try a few surface level bus lines to increase the traffic to the network as a whole but I really don't have alot of understanding of the way things work now.  I guess the wider adoption of private vehicles has reduced traffic on the rail networks, and I know I have a few convoy spacing/density issues on my network, and at this point just breaking even is the first goal, but I just wonder where to start.'


  Overall I think this is one of the most elaborate and impressive Simutrans maps I've ever seen, and there are still unserviced areas to work with, and I feel proper sympathy for the poor server :)
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on April 23, 2012, 01:03:03 PM
thanks dust.

I think 'Turnleigh Brambly Lane Railway Station' or 'Murdingley Market Square Railway Station' will see a lot more traffic once the network is fully up.

You might want to replace some of your high speed lines with maglev. It seems to be quite profitable to connect stations that already have 20k+ arrivals/departures. You're definitely in a better position to link up thurmouth with maglev, but to where, I'm not sure.

The maglev lines need to be really straight. Their corner penalty is pretty significant (220km/h).


PS. could you please remove 'Starmead Mint Lane Stop'?
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on April 25, 2012, 08:15:10 PM
thanks dust.

PEX company is now: Power Gen. Co.
Password is: password
I'm opening it, because the roster is full and I've taken over some of the mail it used to do.


I need to restructure the mail service in the north west so that it uses the hub-spoke layout. The previous method had too many bottlenecks and no amount of road service could alleviate it.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on May 08, 2012, 02:37:28 PM
James, I think you should reset the password for player 4 and 6. I would also liquidate md, who only has $2700 in the bank, which is enough to do nothing.

Not sure about player 2, rebel logistics, sweions regional and abellio. Those companies are not doing anything in particular, but do have plenty of money in the bank. It would be nice if someone can put it to good use, like the former PEX company.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: dustNbone on May 09, 2012, 11:00:40 PM
AEO, I'd like to provide some mail service in the eastern regions, mainly by road, would like a good place to be able to pull it from (maybe a rail platform at one of your PEX stations on the east coast.  Not sure you've been handling mail more so you'll have a good idea of the volume involved.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on May 10, 2012, 12:14:29 AM
AEO, I'd like to provide some mail service in the eastern regions, mainly by road, would like a good place to be able to pull it from (maybe a rail platform at one of your PEX stations on the east coast.  Not sure you've been handling mail more so you'll have a good idea of the volume involved.

Mail works better if you layout the lines in a hub and spoke configuration, similar to how the middle section is on the west coast. I've been meaning to overhaul the west coast, but I can't do it right now.


You can use the PEX lines if you want.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on May 25, 2012, 08:40:55 PM
currently, in 2057, March. Server and clients seem to crash as soon as the game loads.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on May 25, 2012, 08:50:43 PM
Ahh, thank you for the report. Cannot do anything at present, unfortunately, as I am currently on holiday. Will investigate on my return!
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on May 29, 2012, 01:29:17 PM
Thank you for the report - I have tracked down the bug (I think) and fixed it on the -devel branch.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: dustNbone on June 07, 2012, 12:47:13 AM
Was wondering if the fixed devel version is currently running on the server, I try to connect with it and it shows a PakSet mismatch.  Not sure if this would be caused by a mismatch in game version or not.  Thanks for the fix by the way James :)
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: rsdworker on June 07, 2012, 12:49:53 AM
i think james has updated it but forget to apply new patch for us to make connection
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on June 07, 2012, 08:06:26 PM
Ahh, no, the -devel version is not running on the server. The -devel version is the candidate for the next major release, the deployment of which has been delayed somewhat due to waiting for Bernd to finish fixing the physics issues, the fixing of which should hopefully lead to a substantially improved accuracy of the physics calibrations. The new version will also introduce realistic physics for breaking.

Whilst I wait for those issues to be addressed, I have been working on the pakset, which strikes me as a useful use of time, as there is much pakset work to be done.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: diegoviagens on July 04, 2012, 05:22:04 PM
I'd like to play.
I await my password
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on July 04, 2012, 05:37:28 PM
Thank you for your interest in this. The server is down at present as a result of a bug in the current release version. The next version, which fixes the bug, is currently being tested. The server sould be back up again when the new version is released.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: AP on October 07, 2012, 04:56:54 PM
Hi James Just curious, is the server up or not? It's showing up in the in-game list as available, but my game crashes if I try to load it..
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on October 07, 2012, 10:18:31 PM
It's not running properly yet: I need to upgrade to the next version of Simutrans-Experimental and of the pakset for this to be runnable, but I've been exceedingly busy these last few weeks, so haven't had the time to do things like recalibrate the physics of all the steam locomotives as I need to do before I can release the next pakset version. I shall get around to it eventually! Sorry to keep you all waiting.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on October 08, 2012, 12:03:50 AM
I certainly look forward to the new version! :)
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on October 08, 2012, 12:46:49 AM
One thing that might help to speed along the progress is if anybody could assist with integrating/merging the new bridges from the Standard version of the pakset. Currently, there are the bridges from Standard without weight limits alongside the old bridges from Experimental without separate graphics for each type, and, further, the new Standard types are not all equivalent to the old Experimental types, and the prices of both types are too low. If anyone could propose and write the details for a coherent scheme for integrating the both, discarding the superfluous (without discarding any of the lovely new graphics), and sensible compat.tab settings to enable loading of older saved games, this would be exceedingly helpful.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: ӔO on October 09, 2012, 09:23:33 AM
I think I will take a stab at those bridges.
Title: Re: New Simutrans-Experimental server (bridgewater-brunel.me.uk)
Post by: jamespetts on October 09, 2012, 09:34:31 AM
Thank you! That is very kind and would greatly assist in expediting matters.