Servers not being controlled from panel

[AccelerateServers]

We have had a few of our users mention that they are unable to update their servers and upon investigation, when stopping their server using the panel, the servers are still running. I used HLSW/game client to connect to the server to verify it is still running, despite panel saying it is stopped. This is a problem with updating, since the panel triggers the update but the server is still running. I can work around the issue by stopping the server with the panel then using "rcon quit" from RCON. When I start the server using the panel, it seems to start without issue. This is happened one too many times now so I am reporting the issue with a hopeful fix.

.783
Ubuntu 12.04

[Masher]

Please check your log files (both the web log tool and the actual file logs) for any errors. Post any errors if you could.

[AccelerateServers]

I don't have much, this is what is logged when the automatic update occurred last night:

Code: Select all

8/19/2015 8:02:58 PM    Starting automatic update of ID: 37, appid=440.  We have ver 2926419 and required version is 2927789
8/19/2015 8:02:58 PM    Shutting down server ID: 37 for automatic update.
8/19/2015 8:02:58 PM    Shutting down server ID 37
8/19/2015 8:05:05 PM    Error sending update started report
Message could not be sent.
8/19/2015 8:05:17 PM    Update finished for server id: 37 on 8/19/2015 8:05:17 PM
8/19/2015 8:05:34 PM    Attempting to start id:  37...SUCCESS

and then at the end of the log where I started it:

Code: Select all

8/20/2015 5:19:10 PM    Shutting down server ID 37
8/20/2015 5:19:11 PM    Remote Stop Request Received for id: 37 SUCCESS
8/20/2015 5:19:14 PM    Remote Update Received for id: 37
8/20/2015 5:20:46 PM    Remote Update Complete for id: 37
8/20/2015 5:21:03 PM    Attempting to start id:  37...SUCCESS
8/20/2015 7:02:00 PM    Shutting down server ID 37
8/20/2015 7:02:00 PM    Remote Stop Request Received for id: 37 SUCCESS
8/20/2015 7:02:47 PM    Remote Start Request Received for id: 37 SUCCESS

[Masher]

The error sending update report is just stating it was unable to send the update emails. You may just not have email setup.

Are you using the -autoudpate command line option on any of the servers? You'll want to remove those. Maybe also add -norestart to prevent the srcds shell script from trying to restart them.

[AccelerateServers]

I will try the -norestart option and report back.

[AccelerateServers]

Is there any reason to NOT have -norestart in the launch options? Does it affect the way a server is restarted after crash? Or does the panel handle that?

[Masher]

You'll probably want to continue to use -norestart (is this a newish feature)? The panel will restart crashed servers if you have monitoring enabled for the server. Probably explains your issue. You want the panel to manage starts/stops. If something is started outside of the panel the panel has no way of knowing it.

You can also add -strictportbind which will cause a server to fail to start if it can't bind to the specified ports. Which should prevent having extra copies of the servers running if a server is already using those ports.

[AccelerateServers]

It appears that -norestart causes the server to not boot.

[Masher]

Yes, sorry for that; it doesn't work.

I'm tested various things and can't reproduce the original issue. Can you make a backup (use the import/backup admin tool) of one of your servers and email it to me?

[AccelerateServers]

I'm going to attempt to edit the srcds_run script to not restart the server on crash, ill report back if im able to fix it and ill upload the file i edited for future use.

[AccelerateServers]

BTW, -norestart is actually a valid command as specified inside of srcds_run... which is odd.