Обсуждение: Many processes blocked at ProcArrayLock

Test configuration:
Hardware:
    4P intel server, 60 core 120 hard thread.
    Memory:512G
    SSD:2.4T

PG:
max_connections = 160            # (change requires restart)
shared_buffers = 32GB
work_mem = 128MB
maintenance_work_mem = 32MB
bgwriter_delay = 100ms            # 10-10000ms between rounds
bgwriter_lru_maxpages = 200        # 0-1000 max buffers written/round
bgwriter_lru_multiplier = 2.0        # 0-10.0 multipler on buffers scanned/round
wal_level = minimal            # minimal, archive, or hot_standby
wal_buffers = 256MB            # min 32kB, -1 sets based on shared_buffers
autovacuum = off
checkpoint_timeout=60min
checkpoint_segments = 1000
archive_mode = off
synchronous_commit = off
fsync = off
full_page_writes = off


We use tpcc and pgbench to test postgresql 9.4beat2 performance. And we found the tps/tpmc could not increase with the
terminalincrease. The detail information is in attachment. 

Many processes is blocked, I dump the call stack, and found these processes is blocked at: ProcArrayLock. 60% processes
isblocked in ProcArrayEndTransaction with ProcArrayLock EXCLUSIVE, 20% is in GetSnapshotData with ProcArrayLock SHARED.
Otherslocks like XLogFlush and WALInsertLock are not very heavy. 

Is there any way we solve this problem?

On Tue, Dec 2, 2014 at 5:07 PM, Xiaoyulei <xiaoyulei@huawei.com> wrote:
> Test configuration:
> Hardware:
>         4P intel server, 60 core 120 hard thread.
>         Memory:512G
>         SSD:2.4T
>
> PG:
> max_connections = 160                   # (change requires restart)
> shared_buffers = 32GB
> work_mem = 128MB
> maintenance_work_mem = 32MB
> bgwriter_delay = 100ms                  # 10-10000ms between rounds
> bgwriter_lru_maxpages = 200             # 0-1000 max buffers written/round
> bgwriter_lru_multiplier = 2.0           # 0-10.0 multipler on buffers scanned/round
> wal_level = minimal                     # minimal, archive, or hot_standby
> wal_buffers = 256MB                     # min 32kB, -1 sets based on shared_buffers
> autovacuum = off
> checkpoint_timeout=60min
> checkpoint_segments = 1000
> archive_mode = off
> synchronous_commit = off
> fsync = off
> full_page_writes = off
>
>
> We use tpcc and pgbench to test postgresql 9.4beat2 performance. And we found the tps/tpmc could not increase with
theterminal increase. The detail information is in attachment. 
>
> Many processes is blocked, I dump the call stack, and found these processes is blocked at: ProcArrayLock. 60%
processesis blocked in ProcArrayEndTransaction with ProcArrayLock EXCLUSIVE, 20% is in GetSnapshotData with
ProcArrayLockSHARED. Others locks like XLogFlush and WALInsertLock are not very heavy. 
>
> Is there any way we solve this problem?
Providing complete backtraces showing in which code paths those
processes are blocked would help better in understand what may be
going on.
--
Michael

I put all the stack in attachment.

> -----Original Message-----
> From: Michael Paquier [mailto:michael.paquier@gmail.com]
> Sent: Wednesday, December 03, 2014 8:48 AM
> To: Xiaoyulei
> Cc: pgsql-hackers@postgresql.org
> Subject: Re: [HACKERS] Many processes blocked at ProcArrayLock
>
> On Tue, Dec 2, 2014 at 5:07 PM, Xiaoyulei <xiaoyulei@huawei.com> wrote:
> > Test configuration:
> > Hardware:
> >         4P intel server, 60 core 120 hard thread.
> >         Memory:512G
> >         SSD:2.4T
> >
> > PG:
> > max_connections = 160                   # (change requires restart)
> > shared_buffers = 32GB
> > work_mem = 128MB
> > maintenance_work_mem = 32MB
> > bgwriter_delay = 100ms                  # 10-10000ms between rounds
> > bgwriter_lru_maxpages = 200             # 0-1000 max buffers
> written/round
> > bgwriter_lru_multiplier = 2.0           # 0-10.0 multipler on buffers
> scanned/round
> > wal_level = minimal                     # minimal, archive, or
> hot_standby
> > wal_buffers = 256MB                     # min 32kB, -1 sets based on
> shared_buffers
> > autovacuum = off
> > checkpoint_timeout=60min
> > checkpoint_segments = 1000
> > archive_mode = off
> > synchronous_commit = off
> > fsync = off
> > full_page_writes = off
> >
> >
> > We use tpcc and pgbench to test postgresql 9.4beat2 performance. And we
> found the tps/tpmc could not increase with the terminal increase. The detail
> information is in attachment.
> >
> > Many processes is blocked, I dump the call stack, and found these processes
> is blocked at: ProcArrayLock. 60% processes is blocked in
> ProcArrayEndTransaction with ProcArrayLock EXCLUSIVE, 20% is in
> GetSnapshotData with ProcArrayLock SHARED. Others locks like XLogFlush and
> WALInsertLock are not very heavy.
> >
> > Is there any way we solve this problem?
> Providing complete backtraces showing in which code paths those processes
> are blocked would help better in understand what may be going on.
> --
> Michael

Xiaoyulei wrote:
> I put all the stack in attachment.

Not sure that this is really all that useful.  At least I don't have the
patience to examine all this, and I'm not sure it contains the needed
info in the first place.  If you were to ensure your build is using
-fno-omit-frame-pointer in cflags and then used "perf record -a -g"
while the test runs and "perf report -g" once it's finished, you'd get a
useful profile that would show who is acquiring the problematic lock and
why.

-- 
Álvaro Herrera                http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services