Though I thought this might be related to
https://bugzilla.icculus.org/show_bug.cgi?id=1788 it doesn't seem to be.
The amd64 dedicated server will hang within a day or so of running. It'll
either crash out and restart, use 0 CPU and reject all connections, or it'll use
a lot of CPU doing seemingly nothing and also reject all connections.
Not stable. Is there anyway to get a debugging trace to send you?
(Please note that this hasn't ever been tried, but the source suggests it _should_ work)
Send the server a SIGBUS:
kill -s BUS process_number
This _should_ make the server think it segfaulted, and write out a meaningful stack trace for whatever it
is doing at the time. It'll be in the logfile.
(other signals, like SIGSEGV, might do it, too, but some aren't handled as "crash" signals by the engine)
--ryan.
(Please note that this hasn't ever been tried, but the source suggests it _should_ work) Send the server a SIGBUS: kill -s BUS process_number This _should_ make the server think it segfaulted, and write out a meaningful stack trace for whatever it is doing at the time. It'll be in the logfile. (other signals, like SIGSEGV, might do it, too, but some aren't handled as "crash" signals by the engine) --ryan.