Other informational details about configurations are:
shared_buffers = 80GB # min 128kB
# (change requires restart)
temp_buffers = 2GB # min 800kB
work_mem = 2GB # min 64kB
maintenance_work_mem = 16GB # min 1MB
checkpoint_segments = 64 # in logfile segments, min 1, 16MB each
#wal_buffers = -1 # min 32kB, -1 sets based on shared_buffers
#checkpoint_timeout = 5min # range 30s-1h
#checkpoint_completion_target = 0.5 # checkpoint target duration, 0.0 - 1.0
#checkpoint_warning = 30s # 0 disables
effective_cache_size = 130GB
Let me know if you need more information.
On Mon, Aug 25, 2014 at 10:04 AM, Dhruv Shukla <dhruvshukla82@xxxxxxxxx> wrote:
We have a script which notifies us if there are any locks that its waiting for and it sends an email out. But nothing has came thru ...here is one pid which has got stuck from midnight and is still running, but doesnt seem like it will finish anytime soon. It was run manually by me using an strace.PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND64144 postgres 20 0 81.2g 4.1g 4.1g R 100.0 1.1 537:59.86 postmaster
I even checked the lock scripts by running each of the query but nothing has came thru.https://wiki.postgresql.org/wiki/Lock_MonitoringCent OS 6.3, Postgres 9.3.3, Postgis.On Mon, Aug 25, 2014 at 1:54 AM, Albe Laurenz <laurenz.albe@xxxxxxxxxx> wrote:
Dhruv Shukla wrote:That sounds a lot like locking problems.
> Right now we are doing is only logs where on each statement for a sql function we are raising an info.
>
> This way we know which step its getting stuck. Today it got stuck on the insert statement.
>
> And we ran the same SQL function after killing the process, it ran normal without any issues.
Examine pg_locks when a process gets stuck an see if it is waiting for a lock
and who is holding the lock.
Yours,
Laurenz Albe
--
Regards
Dhruv
404-551-2578
Regards
Dhruv
404-551-2578
