Re: Backends stuck in LISTEN
От | Alvaro Herrera |
---|---|
Тема | Re: Backends stuck in LISTEN |
Дата | |
Msg-id | 20140108140058.GH6840@eldon.alvh.no-ip.org обсуждение исходный текст |
Ответ на | Backends stuck in LISTEN (Marko Tiikkaja <marko@joh.to>) |
Ответы |
Re: Backends stuck in LISTEN
|
Список | pgsql-bugs |
Marko Tiikkaja wrote: > Hi people, > > Today we had a problem where around 30-40 backends were stuck in > LISTEN for a long time. I snatched a core dump before we had to > kill the server, and the backtrace looks like this: > > #0 0x00007f26dd17d5b7 in monstartup () from /lib/x86_64-linux-gnu/libc.so.6 > #1 0x00007f26d79e5c70 in ?? () > #2 0x00007f26d79e5c80 in ?? () > #3 0x00007f254d772b60 in ?? () > #4 0x00000000006359c5 in LWLockAcquire (lockid=87, mode=LW_SHARED) > at /build/buildd/postgresql-9.1-9.1.9/build/../src/backend/storage/lmgr/lwlock.c:464 Hm, looks like either they got stuck trying to activate profiling (but then Debian doesn't build its packages with gcc -pg, does it?), or the backtrace is flawed for those upper frames. Maybe installing the libc -dbg (I imagine it'd be libc6-dbg) package would be useful, if this reoccurs. -- Álvaro Herrera http://www.2ndQuadrant.com/ PostgreSQL Development, 24x7 Support, Training & Services
В списке pgsql-bugs по дате отправления: