BYOND Version:506
Operating System:Linux
Web Browser:Firefox 27.0
Applies to:Dream Daemon
Status: Open

Issue hasn't been assigned a status value.
I am trying to fit the crash report, plus a side-error onto one bug report, the BUG messages are the side stuff.

Descriptive Problem Summary:
Server crashed, some libpng error. I am also seeing (For some reason) a big boost in regular error messages here are some form today's log prior to the crash:

BUG: Sequence number 5BF expected but 43 received
BUG: Sequence number 3AE4 expected but 375B received
BYOND Error: failed to certify *Someome*
BUG: Failed to decode message 129,36
BUG: Network connection for *Someone2* shutting down due to read error. (0,0)
BUG: Failed to decode message 129,25

Any way, crashed today because of some libpng error, I am unsure what or why it was about. But it happened.
Here is the crashes logfile:

And here is my core file, it is 600 MB big, hope that's not too much. (Compressed, of course.) 0B1UXNp_ofzs_SGJrODc3dVc3eHM/edit?usp=sharing

Did the problem NOT occur in any earlier versions? If so, what was the last version that worked?

I would guess non-threaded mode worked better, or stable.
I am currently going to test with map-threads off, and see if this performs any different.

Second crash occurred, this time with map-threads off, normal threads on. the .log file remained empty this time, I do have a core file, however. This one is quite smaller than the first.

LINK: 0B1UXNp_ofzs_U1R0UXBtRFJZem8/edit?usp=sharing
The network errors I would not expect to be coming from threads; I simply think you're getting bad data from some users or from your connection. Can't rule out a bug obviously, but these kinds of errors have appeared in the past without threading, and they only come up in the event of data being incorrect.

Your log suggests that libpng had trouble reading a file because it was corrupt, but that it didn't cause the crash. I'll trace the crash to see where that comes out.
They were a side-issue, I haven't seen any of these until in the last few weeks, and was unsure if there was a way to deal with them by myself.
On an unrelated note (and perhaps you can followup in the other thread)-- have you noticed any improvement since running the threading on the latest build? We are desperately trying to figure out if it makes a difference at all, for all the work that's gone into it.
I haven't had a lot of time to test it out, mainly because of sleep problems, and obviously, the crashes.

But other than that, I had some CPU spikes from time to time, which made me feel like I was teleporting a couple of tiles forward from time to time. But I am going to blame that on the bad code that was running (And was on top on the profiler).

But probably because of that laggy proc, I wasn't able to completely feel a lot of difference.

I will however, as soon as the crashes are less and stability is more, start using full threading, though.