I've yet to determine what the actual issue is. But after a round restart in SS13, DreamDaemon locks up and holds at 100% CPU usage before crashing.
It did it again. But instead of crashing, it simply hung. I was able to pull this from the logs.
BUG: Failed to fseek() in ClientSocket::SendNextPacket(159,0,0,1,993): Illegal seek
BUG: Failed to send 'resources'
This repeated about 10 or so times. Any insight?
Its running through BYOND Panel. Which, if I'm correct, runs a VM through linux. The crashing happens every other round. It'll lock up and either crash or continue. After which it crashes a bit later after continuing.
| |
What BYOND version are you using?
Have you tried upgrading or downgrading? | |
Current version is beta 507.1273
| |
Try downgrading to a stable version such as 506.1247, see if that helps at all.
| |
The server provider says its the code. But I see nothing in the logs even hinting at a crash.
| |
This was prior to the most recent crash.
BUG: Sequence number 5218 expected but 42 received BUG: Failed to decode message 2,10 BUG: Network connection for Marvy shutting down due to read error. (1,1) BUG: Failed to decode message 129,34 BUG: Network connection for Teylor shutting down due to read error. (1,1) BUG: Sequence number F352 expected but 81 received | |
Might want to submit a bug report then, Lummox may be able to help you out, especially since you've got the logs.
| |
Who is Lummox?
Also I've downgraded to the latest build. I'll report back after. We had around 35 players. Blobs round and all. Everything went smooth. But upon round end, the CPU usage spiked to 100% before going blank and failing to reset forcing a manual shutdown and reboot. | |
JayEhh wrote:
Who is Lummox? He is a BYOND developer. Also I've downgraded to the latest build. I'll report back after. Great. We had around 35 players. Blobs round and all. Everything went smooth. But upon round end, the CPU usage spiked to 100% before going blank and failing to reset forcing a manual shutdown and reboot. I've seen this happen so many times it just sits and hangs at 100% and can do for hours .. and I'm never sure if it's DreamDaemon or if it's poor programming .. | |
Crashed again. Managed to go nearly 24 hours without a crash before locking up about 2 hours into a round.
| |
Does anyone think having an SQL database linked to the server could be the cause?
| |
No unless you notice your save system is not working properly
| |
The crashing is just random with nothing in the crash logs
| |
JayEhh wrote:
The crashing is just random with nothing in the crash logs If your the client I think you are, you have a 300+ MB core file and a large log file with tons of errors.. Like I said above .. raise a bug report and supply those files so someone can actually help you.. | |
Do you mean tgstation.log? Because I've already checked that.
| |
Why don't you go ahead and post it, or post a link to it.
Along with the core file, and make a bug report. | |
There.
http://www.byond.com/forum/?post=1762050 I posted. If I can't get any reason behind such I don't know what else to do. | |
I have Away Mission delay set to 18800. Wouldn't that be enough to prevent it from loading?
| |
If not, maybe try to see what happens if you kill certain connections, one by one, when it happens (see if that un-freezes it). If it's repeatedly trying to do that operation on a client that for some reason chokes it (ie it has no limit on how many times it'll try that over and over), then removing that connection should let it recover.
Also, are you running daemon on a completely dedicated server, ie no other instances of daemon or even dream seeker (for example some people try to run servers from computers they also play on). If it's not, then see what happens if you kill other processes, especially seeker.
Of course, that depends on timing: if it's crashing after only 2 seconds of stall, there's not much of a window to do the above.
Finally, if you're on windows, see if the problem comes up when running the server on linux (and vice versa).