helling3r 10 Posted March 16, 2011 Hello, yesterday i got an issue with my linux server. After i installed some new missions (total in mpmissions about 40) the server crashed at the moment anybody joined (client said "session lost", server report said "could not load <some_mission_file>.pbo"). The mission file is fine and also the config is not the issue. Somewhere at the forums i read about an ulimit problem and so i turned up the open-filehandle limit as suggested there. After that, players where able to join, but after connecting (the server report says this) the client hangs with "waiting for hosts". This black screen does not go away, not even after 10 minutes. Also the server report does not log anything else. Just as a test i removed some of the mission files, now i only have about 6 or 8 (containing the one formerly reported as "could not load") and all works normal as expected. Is this a known problem and how can i debug that? ---------- Post added at 03:38 PM ---------- Previous post was at 03:37 PM ---------- Oh, forget to mention: Its the recent Arma2-OA, combined operations (arma2 loaded with -mod...) Share this post Link to post Share on other sites
Lonestar 11 Posted March 16, 2011 One common Linux problem is that mission names must be in lower case. Use the ./tolower command for that purpose. Share this post Link to post Share on other sites
sickboy 13 Posted March 16, 2011 Perhaps related to http://www.armastack.info/index.php/910/does-linux-arma-server-crash-when-addons-mpmissions-are-used if not the tolower. Share this post Link to post Share on other sites
Riffler 10 Posted March 16, 2011 (edited) 1 step: use tolower 2 step: check your server.cfg correct code: // MISSIONS CYCLE //class Missions {} class Missions { class WarfareV2_067LiteCO_Chernarus { template = "WarfareV2_067LiteCO.Chernarus"; ...some other code wrong code: // MISSIONS CYCLE //class Missions {} class Missions { class WarfareV2_067LiteCO[color="Red"].[/color]Chernarus { template = "WarfareV2_067LiteCO.Chernarus[color="Red"].pbo[/color]"; ...some other code greetz from germany Edited March 16, 2011 by Riffler Share this post Link to post Share on other sites
UnNamedRUS 10 Posted March 16, 2011 I'm sorry for my english. Also faced with the problem on the Linux server, when the players "waiting for host". It was found that the problem was the MTU setting on the server. Many players use in their own ISP connection pptp. In the end, putting on the server MTU = 1492, part of the players could enter the server. For some players had to prescribe the MTU that is suitable. Here we left a little letter about it. http://dev-heaven.net/issues/18007 Share this post Link to post Share on other sites
icebreakr 3159 Posted March 16, 2011 We've given up on our unix dedicated... too many RPT filling errors :/ Share this post Link to post Share on other sites
helling3r 10 Posted March 17, 2011 As said above, it is not the config. The same, unaltered config runs without problem if i reducr the file number in mpmissions. Of course i have taken care for the mission names to be lowecarse, however thank you for the reminder, it is surely useful to point that out here, if the post is found by others. The tickets[1][2] #11584 and #17018 also contain these tips but as i wrote in my first post, i already rised the limit. My problem is also not the "crash" problem, the server seems to be hanging, its not crashing since i raised the limits. I raised the limits a little more (100 000 open files) and will test if maybe my limits where a little too low (was on soft=10 000, hard=24 000). We've given up on our unix dedicated... too many RPT filling errors :/ Well our linux arma server runs just fine and very smooth. No problems so far, besides when deploying too many missions currently. _________ [1] http://dev-heaven.net/issues/11584 [2] http://dev-heaven.net/issues/17018#note-17 Share this post Link to post Share on other sites
helling3r 10 Posted March 31, 2011 Ok, fixed it: - Indeed one needs to make an entry in the /etc/security/limits.conf - But also, the limits conf is not interpreted in any case; PAM needs to be configured to set limits for a specific context. - in my case (debian lenny) i had to add a line to /etc/pam.d/su: "session required pam_limits.so" Share this post Link to post Share on other sites
Tankbuster 1747 Posted March 31, 2011 We've given up on our unix dedicated... too many RPT filling errors :/ Me too. Share this post Link to post Share on other sites