Process hangs with 100% CPU after several weeks uptime

aio-libs / aiozmq

Asyncio (pep 3156) integration with ZeroMQ

aiozmq.readthedocs.org

BSD 2-Clause "Simplified" License

422 stars 56 forks source link

Process hangs with 100% CPU after several weeks uptime #160

Closed diefans closed 4 years ago

diefans commented 4 years ago

I have a zmq worker hanging after several weeks running fine:

# strace -p 32709 -f 
strace: Process 32709 attached with 4 threads
strace: [ Process PID=32709 runs in x32 mode. ]
[pid 32712] epoll_pwait(15,  <unfinished ...>
[pid 32711] epoll_pwait(12,  <unfinished ...>
[pid 32710] futex(0x234d9e0, FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, 0, NULL, 0xffffffff

If anybody has a hint howto proceed, let me know....

JelleZijlstra commented 4 years ago

Seems like the fourth (main) thread is not doing any syscalls. You could try running gdb -p 32709 and running bt to figure out what it's doing.

diefans commented 4 years ago

ok found the bug... sorry for the noise... nothing to see here ;) btw. https://github.com/benfred/py-spy is very usefull