Re: [hlds_linux] Is there any information we can provide to Valve to help resolve the recent server performance issues?
Fletcher mentioned in one of the earlier e-mails that they are aware of it. We have it too, with net_graph increase hickup/higher times. The graph isn't the same as before and multiple people on our servers are complaining about lag too (but that's usually default, but now the regulars are saying it) -Original Message- From: hlds_linux-boun...@list.valvesoftware.com [mailto:hlds_linux-boun...@list.valvesoftware.com] On Behalf Of Essay Tew Phaun Sent: maandag 29 oktober 2012 3:26 To: Half-Life dedicated Linux server mailing list Subject: [hlds_linux] Is there any information we can provide to Valve to help resolve the recent server performance issues? We've noticed our servers acting very poorly after the Halloween update. Our servers were pretty smooth prior, now players are reporting a lot of random freezing issues and I'm noticing a lot of gaps in my own net_graph report. I did not observe any of this prior to the Halloween update. Both of our systems run CentOS. One is 6.2 64bit and the other 6.3 32bit. Are there any details we can further provide to help rectify this issue? http://i.imgur.com/f7UnP.png ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux
Re: [hlds_linux] Is there any information we can provide to Valve to help resolve the recent server performance issues?
Getting the same freezing issues. Running gentoo, kernel 2.6.31 on a quad core Xeon X3460. 32 player server, tried with and without sourcemod, the same. There were no freezing issues before the update. The freezes occur every ~20-40 seconds and teleports you back a few meters... On 2012.10.29. 4:26, Essay Tew Phaun wrote: We've noticed our servers acting very poorly after the Halloween update. Our servers were pretty smooth prior, now players are reporting a lot of random freezing issues and I'm noticing a lot of gaps in my own net_graph report. I did not observe any of this prior to the Halloween update. Both of our systems run CentOS. One is 6.2 64bit and the other 6.3 32bit. Are there any details we can further provide to help rectify this issue? http://i.imgur.com/f7UnP.png ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux
Re: [hlds_linux] Trends
http://www.teamfortress.com/post.php?id=2338 cl_gameserver_list You will see Standing Trend. Standing = Good/Bad Trend = Downward Fast, Steady, Upward Fast You obviously want your server to be in Good standing, and ideally trending Steady or Upward Fast. On 29/10/2012 6:22 AM, Benedict Glover wrote: What do the different trends mean? Steady, Downward Fast, Upward Fast - which is better, which will get more people? ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux
Re: [hlds_linux] Method To Increase Server Traffic
All of Cameron Munroe's suggestions have been implemented. There is now a 256Bit SSL certificate protecting the entire website. Server hosts can win prizes too now. Refer the most players and you win the prize. November's server prize is $250 USD. More info at: http://www.capturetheprize.com/server-op.php ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux
Re: [hlds_linux] Trends
You can't always be going upwards fast of course. My server has a Good standing and goes through an upwards - steady - downwards trendcycle about every few days, depending on the amount of traffic I'm getting. As long as your standing is Good and you're getting quickplay traffic if your server is set to tf_mm_servermode 1 (or 2 for MvM) you shouldn't have to worry about anything. On Linux you can grep your logfiles to see if you have the same cycle of server trends as I do. On Mon, Oct 29, 2012 at 8:16 AM, Yun Huang Yong gumby_li...@mooh.org wrote: http://www.teamfortress.com/post.php?id=2338 cl_gameserver_list You will see Standing Trend. Standing = Good/Bad Trend = Downward Fast, Steady, Upward Fast You obviously want your server to be in Good standing, and ideally trending Steady or Upward Fast. On 29/10/2012 6:22 AM, Benedict Glover wrote: What do the different trends mean? Steady, Downward Fast, Upward Fast - which is better, which will get more people? ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux -- Idleness is not doing nothing. Idleness is being free to do anything. - Floyd Dell ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux
Re: [hlds_linux] Server Crashing / Freezing - DataTable warning console floods
I can confirm this on a vanilla 'eventmix' halloweenserver running on Ubuntu server 10.04 LTS, and also some variations on the error: DataTable warning: (class prop_dynamic): Out-of-range value (-.x) in SendPropFloat 'm_vNormal', clamping. DataTable warning: prop_dynamic: Out-of-range value (.x) in SendPropFloat 'm_vNormal', clamping. I've noticed that the first with (class prop_dynamic) always has a negative value, the second slightly different error always has a positive value. Hopefully this means anything to anyone at Valve. It doesn't seem to affect my server in any noticable way though, so for now I'm just treating it as console spam that I can ignore. It's logged in my console.log though if anyone at Valve wants to see it. On Mon, Oct 29, 2012 at 12:41 AM, Marco Padovan e...@evcz.tk wrote: This: DataTable warning: player: Out-of-range value (.x) in SendPropFloat 'm_flLastDamageTime', clamping. is still happening :/ ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux -- Idleness is not doing nothing. Idleness is being free to do anything. - Floyd Dell ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux
Re: [hlds_linux] Server Crashing / Freezing - DataTable warning console floods
In certain cases it reaches very high levels and make your server hang/freeze. This extension cures it: http://forums.alliedmods.net/showthread.php?p=1817962 but hope one day even vanilla servers without sourcemod can be cured with an official fix ;) Il 29/10/2012 09.35, Rudy Bleeker ha scritto: I can confirm this on a vanilla 'eventmix' halloweenserver running on Ubuntu server 10.04 LTS, and also some variations on the error: DataTable warning: (class prop_dynamic): Out-of-range value (-.x) in SendPropFloat 'm_vNormal', clamping. DataTable warning: prop_dynamic: Out-of-range value (.x) in SendPropFloat 'm_vNormal', clamping. I've noticed that the first with (class prop_dynamic) always has a negative value, the second slightly different error always has a positive value. Hopefully this means anything to anyone at Valve. It doesn't seem to affect my server in any noticable way though, so for now I'm just treating it as console spam that I can ignore. It's logged in my console.log though if anyone at Valve wants to see it. On Mon, Oct 29, 2012 at 12:41 AM, Marco Padovan e...@evcz.tk wrote: This: DataTable warning: player: Out-of-range value (.x) in SendPropFloat 'm_flLastDamageTime', clamping. is still happening :/ ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux
Re: [hlds_linux] Optional Team Fortress 2 update released
It might not be, but just to rule it out... If you're getting lag, try cleaning out /orangebox/tf/downloads. If you have replays enabled, also clean out /orangebox/tf/replay/server/sessions and /orangebox/tf/replay/server/blocks, as for some reason those folders aren't cleaned up automatically. When there is too many files in one folder, it seems to cause the server to lag whenever it writes to said overfilled folder. :/ We already had a script to clean out the downloads folder, but overlooked the sessions and blocks folder as we thought it was enough to clean out the folder defined in replay_local_fileserver_path. On Mon, Oct 29, 2012 at 12:26 PM, Peter Reinhold peter_va...@reinhold.dkwrote: On 28.10.2012 01:19, Essay Tew Phaun wrote: I'm having lag on PvP servers, I don't think it's just an MvM thing. Experiencing strange lag-spikes on both regular TF2 and MvM servers since the update /Peter __**_ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.**com/cgi-bin/mailman/listinfo/**hlds_linuxhttps://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux
Re: [hlds_linux] Optional Team Fortress 2 update released
We actually run those scripts daily to clean out replays/downloads and several log folders too. But indeed, before we did that we had lag spikes also because of it. But thats now not the case. 2012/10/29 Emil Larsson ail...@gmail.com It might not be, but just to rule it out... If you're getting lag, try cleaning out /orangebox/tf/downloads. If you have replays enabled, also clean out /orangebox/tf/replay/server/sessions and /orangebox/tf/replay/server/blocks, as for some reason those folders aren't cleaned up automatically. When there is too many files in one folder, it seems to cause the server to lag whenever it writes to said overfilled folder. :/ We already had a script to clean out the downloads folder, but overlooked the sessions and blocks folder as we thought it was enough to clean out the folder defined in replay_local_fileserver_path. On Mon, Oct 29, 2012 at 12:26 PM, Peter Reinhold peter_va...@reinhold.dk wrote: On 28.10.2012 01:19, Essay Tew Phaun wrote: I'm having lag on PvP servers, I don't think it's just an MvM thing. Experiencing strange lag-spikes on both regular TF2 and MvM servers since the update /Peter __**_ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.**com/cgi-bin/mailman/listinfo/**hlds_linux https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux
Re: [hlds_linux] Optional Team Fortress 2 update released
On 29.10.2012 13:41, Erik-jan Riemers wrote: We actually run those scripts daily to clean out replays/downloads and several log folders too. But indeed, before we did that we had lag spikes also because of it. But thats now not the case. As do I, this was mentioned on this list some time ago, and I setup some scripts to do this maintenance daily ever since. /Peter ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux
Re: [hlds_linux] Trends
Especially if your server empties for the night, it might go downward fast and then upward fast every day. If server has players, you are doing fine. Some sort of stats will help you on identifying the players like hlxce. Are they coming back or are there just thousands of players that connect, play some and leave, never coming back. -ics - Alkuperäinen viesti - You can't always be going upwards fast of course. My server has a Good standing and goes through an upwards - steady - downwards trendcycle about every few days, depending on the amount of traffic I'm getting. As long as your standing is Good and you're getting quickplay traffic if your server is set to tf_mm_servermode 1 (or 2 for MvM) you shouldn't have to worry about anything. On Linux you can grep your logfiles to see if you have the same cycle of server trends as I do. On Mon, Oct 29, 2012 at 8:16 AM, Yun Huang Yong gumby_li...@mooh.org wrote: http://www.teamfortress.com/post.php?id=2338 cl_gameserver_list You will see Standing Trend. Standing = Good/Bad Trend = Downward Fast, Steady, Upward Fast You obviously want your server to be in Good standing, and ideally trending Steady or Upward Fast. On 29/10/2012 6:22 AM, Benedict Glover wrote: What do the different trends mean? Steady, Downward Fast, Upward Fast - which is better, which will get more people? ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux -- Idleness is not doing nothing. Idleness is being free to do anything. - Floyd Dell ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux
Re: [hlds_linux] Sourcemod STILL not working for me but it is for everyone else
Loading via gameinfo.txt should still work, but you'll have to copy or symlink addons/metamod/bin/server.so to addons/metamod/bin/server_srv.so On 10/28/2012 1:11 AM, ics wrote: Do you load metamod via gameinfo.txt or vdf? You need vdf since gameinfo style wont load metamod snapshot. -ics - Alkuperäinen viesti - Not working for me. Mm wont load. Ill provide versions soon. I have a linux centos rental and i can relay shtuff to my host if need be. Sincerely, Aaron On Oct 27, 2012 9:32 PM, martin v velt...@gmail.com wrote: If it is working for EVERYONE just fine but not for you it seems that it something on your end. You have a thread on allied forums with an istruction how to fix this. If you can't do this maybe you should delete your server? 2012/10/28 Benedict Glover neobened...@hotmail.co.uk Ugh, I feel unbelievably dumb right now. Forgotten I had only temporarily disabled SELinux and had since rebooted my system when Halloween came around. Ugh. Thank you very much. From: neobened...@hotmail.co.uk To: hlds_linux@list.valvesoftware.com Date: Sun, 28 Oct 2012 00:37:35 +0100 Subject: Re: [hlds_linux] Sourcemod STILL not working for me but it is for everyone else I'll try completely disabling this piece of swear word... From: asher...@gmail.com Date: Sun, 28 Oct 2012 00:35:32 +0100 To: hlds_linux@list.valvesoftware.com Subject: Re: [hlds_linux] Sourcemod STILL not working for me but it is for everyone else Freezes on plugin loading is caused by SELinux being enabled and srcds_linux not being granted execstack and execheap. ~ Their heads are green, and their hands are blue, And they went to sea in a Sieve. - Edward Lear On Sun, Oct 28, 2012 at 12:34 AM, Anders Olsson bran...@hotmail.com wrote: Use latest and it will work! Metamod 1.10.0-devV Sourcemod 1.5.0-dev+3675 -Ursprungligt meddelande- From: Giovanni Harting Sent: Sunday, October 28, 2012 1:31 AM To: Half-Life dedicated Linux server mailing list Subject: Re: [hlds_linux] Sourcemod STILL not working for me but it is for everyone else Witch version do you use? Working is latest sm dev snapshot and metamod 1.10 dev snapshot 2012/10/28 Benedict Glover neobened...@hotmail.co.uk Since Alliedmodders isn't giving me any responce, I'd like to know how everyone has managed to get sourcemod working but me. I've tried everything, but it will always lock up whenever it tries to load any plugin (and all I want atm is the ones that come with it). I've tried recompiling the plugins from the scripting folder, tried the windows plugins, but it keeps hanging on: L 10/28/2012 - 01:27:48: Mapchange to koth_lakeside_event Sourcemod loads fine with no plugins in the plugins folder, however as soon as I put one in and try to load it it just freezes. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux -- Nicholas Hastings AlliedMods.net http://www.alliedmods.net ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux
Re: [hlds_linux] Optional Team Fortress 2 update released
One of my servers is having awful lag spikes, the one that runs eventmix. The other that runs event247 on same machine seems to be ok so i guess it's some more random bug or related to some other maps. -ics 29.10.2012 15:13, Peter Reinhold kirjoitti: On 29.10.2012 13:41, Erik-jan Riemers wrote: We actually run those scripts daily to clean out replays/downloads and several log folders too. But indeed, before we did that we had lag spikes also because of it. But thats now not the case. As do I, this was mentioned on this list some time ago, and I setup some scripts to do this maintenance daily ever since. /Peter ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux
Re: [hlds_linux] Optional Team Fortress 2 update released
On 29.10.2012 17:05, ics wrote: One of my servers is having awful lag spikes, the one that runs eventmix. The other that runs event247 on same machine seems to be ok so i guess it's some more random bug or related to some other maps. We're experiencing the lag on both koth_lakeside_event, badwater and MvM maps /Peter ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux
Re: [hlds_linux] Optional Team Fortress 2 update released
I've noticed occasional lag spikes on precisely the same maps on two of my servers. -J On Oct 29, 2012, at 12:15 PM, Peter Reinhold peter_va...@reinhold.dk wrote: On 29.10.2012 17:05, ics wrote: One of my servers is having awful lag spikes, the one that runs eventmix. The other that runs event247 on same machine seems to be ok so i guess it's some more random bug or related to some other maps. We're experiencing the lag on both koth_lakeside_event, badwater and MvM maps /Peter ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux
Re: [hlds_linux] Optional Team Fortress 2 update released
There has been some spikes on others too but nothing major. Usually they are related to events where players join to the server and something happens at the same time that requires more cpu. I'm not sayin there isn't a problem since so many are experiencing it, just thinking out loud what might be the cause. -ics 29.10.2012 18:15, Peter Reinhold kirjoitti: On 29.10.2012 17:05, ics wrote: One of my servers is having awful lag spikes, the one that runs eventmix. The other that runs event247 on same machine seems to be ok so i guess it's some more random bug or related to some other maps. We're experiencing the lag on both koth_lakeside_event, badwater and MvM maps /Peter ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux
[hlds_linux] TF2: map not changing once mp_timelimit is reached
We are running some public servers with just 1 single map specified. mp_timelimit is set to 60 mp_winlimit is set to 0 when issuing timeleft in console we getthis is the last round... but the map never changes anyway. Is there something Imissed? Basically I want the map to change once 60minutes are passed... even if nobody is connected or playing. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux
Re: [hlds_linux] Optional Team Fortress 2 update released
Our Dustbowl server is experiencing some spikes and lag that we never experienced before this update. It seems all of them are, but I mostly play on that server. Our KOTH server is also doing it. I'm not even sure it's map specific. I just think something about the update has borked something. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux
Re: [hlds_linux] Optional Team Fortress 2 update released
If you look at your console you will see the servers spewing out missing files and other errors. It is a wall of red text. Missing X particle, Missing this action. etc. I don't think it is the servers, rather the clients having a hard time trying to process all the effects. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux
Re: [hlds_linux] Optional Team Fortress 2 update released
I noticed one MVM server was lagging terribly with the new coaltown_event map so I took it out of the list and tested it with a few maps. All the MVM maps were practically unplayable, stuttering every few seconds or so. I tried badwater and still got stuttering, albeit significantly less. On turbine and 2fort I didn't notice any issue. I've left that server shut down for the moment. -J On Oct 29, 2012, at 12:25 PM, Essay Tew Phaun sc2p...@gmail.com wrote: Our Dustbowl server is experiencing some spikes and lag that we never experienced before this update. It seems all of them are, but I mostly play on that server. Our KOTH server is also doing it. I'm not even sure it's map specific. I just think something about the update has borked something. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux
[hlds_linux] Sharing Server Files with Unionfs
Hey all. For a while I've been operating multiple TF2 servers off the same data files, where all each instance has is a separate /cfg and /addons directory. I'm using a program called Unionfs-fuse, which allows you to mount directories in layers over each other. (I'm using the FUSE variant for simplicities sake. The kernelspace version would theoretically have better performance but I haven't noticed any issues that would necessitate it.) The way I've set up my directory mounts, I start with a common data files folder that is essentially a complete source game install. For each server instance, I create a config folder that contains just the server-specific files. So for a tf2 server that would be along the lines of tf-config-1/orangebox/tf/addons and tf-config-1/orangebox/tf/cfg. Then I create an empty directory for the actual server instance. Now I use Unionfs to mount the common directory as readonly and the corresponding config directory as readwrite for each specific instance directory. The way Unionfs works, if the top level directory (config) doesn't contain a file, it will look in the bottom level common directory for it. When the server writes files back to the instance directory that it's run from, they're actually put in the config directory. So after running the server, the config directory will get populated with a bunch of caching and log files and the like. Unionfs-fuse is a filesystem type as recognized by mount, though the package I use also has an actual unionfs-fuse command which is effectively an alias to mount. I use the copy on write option which means that if the server tries to write back to a file which is in the readonly common location, it will first copy it to the readwrite config location. For the purposes of updating, I also created another directory that I use the steam updater on. This directory has the common directory mounted as readwrite initially. I also mount two dummy directories over the addons and cfg paths so that if the updater writes anything to those locations it doesn't get put back into common. For ease of configuration, I put several base configuration files in the common directory's cfg location. For example, I've got a standard-server.cfg file which sets up all the options for my standard servers. I've also got mvm-server.cfg which sets up an MvM server. Then in each server-specific config's cfg directory I've got a server.cfg that execs one of the common cfgs and then sets up that particular server's specific options. In practice, that's usually just the quickplay identity and the hostname. Although it makes it easy for me to just tweak a couple small options. E.g. for the halloween event I changed the map rotation for a couple of my servers in their specific cfgs. I also have a maplists directory in the common tree that just has a bunch of text files with maplists that can be loaded both via the mapcycle cvar and in the sourcemod plugin configurations. I've written up a few shell scripts that make my server management easier. One of them can either mount or unmount all of the layered filesystems all at once. I call this script from an init script so that the mounts are automatically performed when the system boots. http://pastebin.com/dw6FLhB8 This script is just a shortcut for using steam to update servers. I created it before I was using the shared data mounts as it meant I could just chain all the server folders as parameters and it would update them one at a time. http://pastebin.com/VV6WcmHs This script is a shortcut for updating sourcemod. Due to limitations of sourcemod itself there has to be a separate addons folder for each server. Therefore I created this script to untar the downloaded package and copy some of the contents to a game server's addons directory. It leaves the configs directory alone. http://pastebin.com/tBzqwZQW So I've been using this system for several months with as many as 10 servers sharing one common data directory. It saves a TON of disk space (each config folder ends up being about 500 MB with all the cached files) and I'm considering replacing the physical disk in one of my servers with an SSD to see how it performs. I also only have to run the updater once to get all the servers on a box up to date. Adding a new server is just a matter of making a new config and instance directory and adding its name to the mount script. Feel free to ask me any questions. Cheers! -Jeremy ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux
Re: [hlds_linux] Optional Team Fortress 2 update released
Are those of you with lag issues watching your CPU utilization? I had similar symptoms on my servers after the pyro update (which lasted for months). The in game lag spikes I saw coincided with srcds CPU spikes. Lord_Jeremy lord.jer...@gmail.com wrote:I noticed one MVM server was lagging terribly with the new coaltown_event map so I took it out of the list and tested it with a few maps. All the MVM maps were practically unplayable, stuttering every few seconds or so. I tried badwater and still got stuttering, albeit significantly less. On turbine and 2fort I didn't notice any issue. I've left that server shut down for the moment. -J On Oct 29, 2012, at 12:25 PM, Essay Tew Phaun sc2p...@gmail.com wrote: Our Dustbowl server is experiencing some spikes and lag that we never experienced before this update. It seems all of them are, but I mostly play on that server. Our KOTH server is also doing it. I'm not even sure it's map specific. I just think something about the update has borked something. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux
Re: [hlds_linux] Sharing Server Files with Unionfs
Nice, thanks for the story. Always good to hear sollutions from people just for the sake of what else is out there Instead of using a uninionfs we use symbolic links, i only share the maps folder because that is one of the biggest ones. I just dump all maps in there regardless if it needs to be there and keep that up to date. Only i use unison to sync from our bigbox to several other servers (physical) , its a bit more space but all the boxes have enough anyways (and the bigbox is ssd only) We link all our plugins to 1 big shared folder (which is synced via unison again) so basicly i can update on any of the servers a plugin in my shared folder and when the sync comes (its in cron) it gets picked up and all the servers using that plugin will be updated too. Updating tf2 itself goes via the SSMS tool (see AM/github), it just sends out a _restart to the server and the srcds script will update because of that when a update comes out. Nothing fancy but it does a good job. Maintenance with 10+ servers is just a horror if you have different mods and such too. Overall its still a lot of work, getting the right people to do the job is even harder. So good idea's are always welcome. 2012/10/29 Lord_Jeremy lord.jer...@gmail.com Hey all. For a while I've been operating multiple TF2 servers off the same data files, where all each instance has is a separate /cfg and /addons directory. I'm using a program called Unionfs-fuse, which allows you to mount directories in layers over each other. (I'm using the FUSE variant for simplicities sake. The kernelspace version would theoretically have better performance but I haven't noticed any issues that would necessitate it.) The way I've set up my directory mounts, I start with a common data files folder that is essentially a complete source game install. For each server instance, I create a config folder that contains just the server-specific files. So for a tf2 server that would be along the lines of tf-config-1/orangebox/tf/addons and tf-config-1/orangebox/tf/cfg. Then I create an empty directory for the actual server instance. Now I use Unionfs to mount the common directory as readonly and the corresponding config directory as readwrite for each specific instance directory. The way Unionfs works, if the top level directory (config) doesn't contain a file, it will look in the bottom level common directory for it. When the server writes files back to the instance directory that it's run from, they're actually put in the config directory. So after running the server, the config directory will get populated with a bunch of caching and log files and the like. Unionfs-fuse is a filesystem type as recognized by mount, though the package I use also has an actual unionfs-fuse command which is effectively an alias to mount. I use the copy on write option which means that if the server tries to write back to a file which is in the readonly common location, it will first copy it to the readwrite config location. For the purposes of updating, I also created another directory that I use the steam updater on. This directory has the common directory mounted as readwrite initially. I also mount two dummy directories over the addons and cfg paths so that if the updater writes anything to those locations it doesn't get put back into common. For ease of configuration, I put several base configuration files in the common directory's cfg location. For example, I've got a standard-server.cfg file which sets up all the options for my standard servers. I've also got mvm-server.cfg which sets up an MvM server. Then in each server-specific config's cfg directory I've got a server.cfg that execs one of the common cfgs and then sets up that particular server's specific options. In practice, that's usually just the quickplay identity and the hostname. Although it makes it easy for me to just tweak a couple small options. E.g. for the halloween event I changed the map rotation for a couple of my servers in their specific cfgs. I also have a maplists directory in the common tree that just has a bunch of text files with maplists that can be loaded both via the mapcycle cvar and in the sourcemod plugin configurations. I've written up a few shell scripts that make my server management easier. One of them can either mount or unmount all of the layered filesystems all at once. I call this script from an init script so that the mounts are automatically performed when the system boots. http://pastebin.com/dw6FLhB8 This script is just a shortcut for using steam to update servers. I created it before I was using the shared data mounts as it meant I could just chain all the server folders as parameters and it would update them one at a time. http://pastebin.com/VV6WcmHs This script is a shortcut for updating sourcemod. Due to limitations of sourcemod itself there has to be a separate addons folder for each server. Therefore I created this script
[hlds_linux] Hanging on mapchange (changelevel)
Hi I'm running these versions of SM and MMS: http://pastebin.com/0sM4p22q My problem is that the server hangs upon mapchange, this is random and I do not know why it happens. I have a stacktrace of it which can be found here: http://pastebin.com/raw.php?i=fLebMQLS Any ideas/suggestions on what to do? ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux
Re: [hlds_linux] Sharing Server Files with Unionfs
Thank you for this tutorial, I am using hard-links, but there are some issues because of server writing to files therefore I had to change the owner of base files to root, but that makes more problems :) I will definitely look at the unionfs. On 2012.10.29. 19:17, Lord_Jeremy wrote: Hey all. For a while I've been operating multiple TF2 servers off the same data files, where all each instance has is a separate /cfg and /addons directory. I'm using a program called Unionfs-fuse, which allows you to mount directories in layers over each other. (I'm using the FUSE variant for simplicities sake. The kernelspace version would theoretically have better performance but I haven't noticed any issues that would necessitate it.) The way I've set up my directory mounts, I start with a common data files folder that is essentially a complete source game install. For each server instance, I create a config folder that contains just the server-specific files. So for a tf2 server that would be along the lines of tf-config-1/orangebox/tf/addons and tf-config-1/orangebox/tf/cfg. Then I create an empty directory for the actual server instance. Now I use Unionfs to mount the common directory as readonly and the corresponding config directory as readwrite for each specific instance directory. The way Unionfs works, if the top level directory (config) doesn't contain a file, it will look in the bottom level common directory for it. When the server writes files back to the instance directory that it's run from, they're actually put in the config directory. So after running the server, the config directory will get populated with a bunch of caching and log files and the like. Unionfs-fuse is a filesystem type as recognized by mount, though the package I use also has an actual unionfs-fuse command which is effectively an alias to mount. I use the copy on write option which means that if the server tries to write back to a file which is in the readonly common location, it will first copy it to the readwrite config location. For the purposes of updating, I also created another directory that I use the steam updater on. This directory has the common directory mounted as readwrite initially. I also mount two dummy directories over the addons and cfg paths so that if the updater writes anything to those locations it doesn't get put back into common. For ease of configuration, I put several base configuration files in the common directory's cfg location. For example, I've got a standard-server.cfg file which sets up all the options for my standard servers. I've also got mvm-server.cfg which sets up an MvM server. Then in each server-specific config's cfg directory I've got a server.cfg that execs one of the common cfgs and then sets up that particular server's specific options. In practice, that's usually just the quickplay identity and the hostname. Although it makes it easy for me to just tweak a couple small options. E.g. for the halloween event I changed the map rotation for a couple of my servers in their specific cfgs. I also have a maplists directory in the common tree that just has a bunch of text files with maplists that can be loaded both via the mapcycle cvar and in the sourcemod plugin configurations. I've written up a few shell scripts that make my server management easier. One of them can either mount or unmount all of the layered filesystems all at once. I call this script from an init script so that the mounts are automatically performed when the system boots. http://pastebin.com/dw6FLhB8 This script is just a shortcut for using steam to update servers. I created it before I was using the shared data mounts as it meant I could just chain all the server folders as parameters and it would update them one at a time. http://pastebin.com/VV6WcmHs This script is a shortcut for updating sourcemod. Due to limitations of sourcemod itself there has to be a separate addons folder for each server. Therefore I created this script to untar the downloaded package and copy some of the contents to a game server's addons directory. It leaves the configs directory alone. http://pastebin.com/tBzqwZQW So I've been using this system for several months with as many as 10 servers sharing one common data directory. It saves a TON of disk space (each config folder ends up being about 500 MB with all the cached files) and I'm considering replacing the physical disk in one of my servers with an SSD to see how it performs. I also only have to run the updater once to get all the servers on a box up to date. Adding a new server is just a matter of making a new config and instance directory and adding its name to the mount script. Feel free to ask me any questions. Cheers! -Jeremy ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux
Re: [hlds_linux] Running a Halloween Server
You seem to be upset with TF2 for requiring a certain amount of people to take on certain tasks. You want to be able to experience 100% of the content regardless of it's intention, delivery, or challenge. This seems like you just want more free stuff for your free game, this is good, it's a sign that you really like this game and you're frustrated you cannot experience everything about it because you must rely on the public to help you. I can understand that's... sometimes unfortunate, but at the same time you cannot hold this against Valve/TF2. I'm not sure what exactly you're expecting for TF2 anymore. I mean think about what you've gotten and how it stacks up to any other multiplayer game you've ever played: - TF2 came out in 2007 (you cite Borderlands 2, when the first Borderlands came out in 2009) In this time there have been: a lot of -free- updates. Over 5 years of constant support, new game modes, new maps, new voice actor lines, new models, new weapons, new hats (hooray??), more features, different entire game modes. All of these things have been provided for free, because some people really like those $100 rings. - TF2 is now -free-, 100% free THERE ARE STILL CONTENT UPDATES. This is what really blows me away. This game is 100% free, this entire event was 100% free. Sure you can say the event sucks but why have any events... ever really? CoD doesn't have Halloween events and it seems to be quite fine without them. I'm not sure Halo has any kind of achievements unlocking new weapons, or new maps coming out (for free). - Not everyone wants to play by themselves. I think the fact that you need at least x players to do something is kind of neat - it's not something you see in a lot of other video games, and I mean is this really a big deal? It's ONE map during ONE week of TF2, and unless you've been sucked into the trading meta-game (that is cheapening TF2) there isn't anything you 'get' out of this. Then again when I started playing TF2 5 years ago, there wasn't anything to get at anytime. You say that 32 man servers show you can take an idea too far, but what is to stop that argument from being flipped on it's head. Why does VALVE get to tell me how many players I should max out on? If one can say 32+ is too many, I'm comfortable in saying you need 6 to do MvM, unless you regularly are able to 1-man MMORPG bosses or L4D runs I don't see why this is such a strange restraint. (Also side note: what is the challenge in hitting the tank? The challenge comes in splitting your attention, the tank itself is easy but it REQUIRES time, REQUIRES damage. If you have to concentrate on stopping a tank or stopping bomb progress what do you do? That question is the reason tanks exist. They are not difficult on normal, but then again the Normal difficulties are pretty easy.) - Perhaps TF2 is reaching the end of it's idea threshold So MvM isn't some perfect MAN VERSUS MACHINE SHOWDOWN, it's more of a really fancy set of maps and AI logic and hacked up missions. I explain MvM as a super polished SourceMod plugin - and that's a blast. It's a shame it takes up a literal 32 man server, but expecting them to completely change bots/AI behavior within their game at this point in it's lifetime just doesn't many any sense - there is no financial motive unless they plan on re-releasing it as some paid addition. You can't just edit SoldierBot.cs and set useRocketJumps = true. After 5 years perhaps TF2 should just be enjoyed for what it is, instead of being reliant on every holiday update bringing with it more fun. Remember how fun cp_dustbowl was on launch day? It's still that fun, you've just come to expect more, a lot more. My advice is: don't cater to quickplayer. They are users that just mashed PLAY NOW and most times have no loyalty to the place they show up at. Take your server off quickplay - get a community of players together that still like TF2 for what it is. If you cannot get those people together, then perhaps your community has moved on from TF2 and Quickplay users were just the band-aid hiding this wound. TF2 is a bunch of fun but it has become a very different game over its lifetime. I agree with your sentiment that the updates now don't feel as fun as they used to, but I also see that I have around 2,000 hours of game time in TF2, and I think that's just a lot of time to spend on one thing in general. On Sun, Oct 28, 2012 at 1:43 AM, dan needa...@ntlworld.com wrote: On 28/10/2012 07:38, Maavrik wrote: I always felt that TF2 was meant to be silly. For instance, the Scout knows he's in a video game, not very serious if you ask me It doesn't follow that in order for the content of the game not to be serious that it doesn't matter what that content of the game is or how the mechanics of it work. If any old crap worked because It's not serious then Splash Damage could be successful PC game developers :) As I said, I think it would be more fun had this particular update considered
Re: [hlds_linux] Running a Halloween Server
Eloquent. On 30/10/2012 7:58 AM, doc drga...@gmail.com wrote: You seem to be upset with TF2 for requiring a certain amount of people to take on certain tasks. You want to be able to experience 100% of the content regardless of it's intention, delivery, or challenge. This seems like you just want more free stuff for your free game, this is good, it's a sign that you really like this game and you're frustrated you cannot experience everything about it because you must rely on the public to help you. I can understand that's... sometimes unfortunate, but at the same time you cannot hold this against Valve/TF2. I'm not sure what exactly you're expecting for TF2 anymore. I mean think about what you've gotten and how it stacks up to any other multiplayer game you've ever played: - TF2 came out in 2007 (you cite Borderlands 2, when the first Borderlands came out in 2009) In this time there have been: a lot of -free- updates. Over 5 years of constant support, new game modes, new maps, new voice actor lines, new models, new weapons, new hats (hooray??), more features, different entire game modes. All of these things have been provided for free, because some people really like those $100 rings. - TF2 is now -free-, 100% free THERE ARE STILL CONTENT UPDATES. This is what really blows me away. This game is 100% free, this entire event was 100% free. Sure you can say the event sucks but why have any events... ever really? CoD doesn't have Halloween events and it seems to be quite fine without them. I'm not sure Halo has any kind of achievements unlocking new weapons, or new maps coming out (for free). - Not everyone wants to play by themselves. I think the fact that you need at least x players to do something is kind of neat - it's not something you see in a lot of other video games, and I mean is this really a big deal? It's ONE map during ONE week of TF2, and unless you've been sucked into the trading meta-game (that is cheapening TF2) there isn't anything you 'get' out of this. Then again when I started playing TF2 5 years ago, there wasn't anything to get at anytime. You say that 32 man servers show you can take an idea too far, but what is to stop that argument from being flipped on it's head. Why does VALVE get to tell me how many players I should max out on? If one can say 32+ is too many, I'm comfortable in saying you need 6 to do MvM, unless you regularly are able to 1-man MMORPG bosses or L4D runs I don't see why this is such a strange restraint. (Also side note: what is the challenge in hitting the tank? The challenge comes in splitting your attention, the tank itself is easy but it REQUIRES time, REQUIRES damage. If you have to concentrate on stopping a tank or stopping bomb progress what do you do? That question is the reason tanks exist. They are not difficult on normal, but then again the Normal difficulties are pretty easy.) - Perhaps TF2 is reaching the end of it's idea threshold So MvM isn't some perfect MAN VERSUS MACHINE SHOWDOWN, it's more of a really fancy set of maps and AI logic and hacked up missions. I explain MvM as a super polished SourceMod plugin - and that's a blast. It's a shame it takes up a literal 32 man server, but expecting them to completely change bots/AI behavior within their game at this point in it's lifetime just doesn't many any sense - there is no financial motive unless they plan on re-releasing it as some paid addition. You can't just edit SoldierBot.cs and set useRocketJumps = true. After 5 years perhaps TF2 should just be enjoyed for what it is, instead of being reliant on every holiday update bringing with it more fun. Remember how fun cp_dustbowl was on launch day? It's still that fun, you've just come to expect more, a lot more. My advice is: don't cater to quickplayer. They are users that just mashed PLAY NOW and most times have no loyalty to the place they show up at. Take your server off quickplay - get a community of players together that still like TF2 for what it is. If you cannot get those people together, then perhaps your community has moved on from TF2 and Quickplay users were just the band-aid hiding this wound. TF2 is a bunch of fun but it has become a very different game over its lifetime. I agree with your sentiment that the updates now don't feel as fun as they used to, but I also see that I have around 2,000 hours of game time in TF2, and I think that's just a lot of time to spend on one thing in general. On Sun, Oct 28, 2012 at 1:43 AM, dan needa...@ntlworld.com wrote: On 28/10/2012 07:38, Maavrik wrote: I always felt that TF2 was meant to be silly. For instance, the Scout knows he's in a video game, not very serious if you ask me It doesn't follow that in order for the content of the game not to be serious that it doesn't matter what that content of the game is or how the mechanics of it work. If any old crap worked because It's not serious
Re: [hlds_linux] TF2: map not changing once mp_timelimit is reached
Looks like it was just a matter of: mp_match_end_at_timelimit :) Il 29/10/2012 17.18, Marco Padovan ha scritto: We are running some public servers with just 1 single map specified. mp_timelimit is set to 60 mp_winlimit is set to 0 when issuing timeleft in console we getthis is the last round... but the map never changes anyway. Is there something Imissed? Basically I want the map to change once 60minutes are passed... even if nobody is connected or playing. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux
Re: [hlds_linux] Invalid STEAM userID ticket, STEAMAUTH Client NAME received failure code 7
Steam hiccup from what I've been told. Nothing you can do about it. On Mon, Oct 29, 2012 at 4:14 PM, bottige...@gmail.com bottige...@gmail.comwrote: Is anyone getting entire servers disconnected with this message in the console: STEAMAUTH Client NAME received failure code 7 And this message on the client? Invalid STEAM userID ticket http://i.imgur.com/6E9OQ.jpg ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux