Re: [PERFORM] postgresql.conf recommendations

2013-02-05 Thread a...@hsk.hk
On 6 Feb 2013, at 12:23 PM, Josh Krupka wrote: > On Tue, Feb 5, 2013 at 6:46 PM, Johnny Tan wrote: > shared_buffers = 48GB # min 128kB > > Hi, From the postgresql.conf, I can see that the shared_buffers is set to 48GB which is not small, it would be possible that the large

Re: [PERFORM] postgresql.conf recommendations

2013-02-05 Thread Pavan Deolasee
On Wed, Feb 6, 2013 at 3:32 AM, Johnny Tan wrote: > > maintenance_work_mem = 24GB # min 1MB I'm quite astonished by this setting. Not that it explains the problem at hand, but I wonder if this is a plain mistake in configuration. Thanks, Pavan -- Pavan Deolasee http://www.linkedin.com/in/pavan

Re: [PERFORM] postgresql.conf recommendations

2013-02-05 Thread Josh Krupka
I've been looking into something on our system that sounds similar to what you're seeing. I'm still researching it, but I'm suspecting the memory compaction that runs as part of transparent huge pages when memory is allocated... yet to be proven. The tunable you mentioned controls the compaction

Re: [PERFORM] postgresql.conf recommendations

2013-02-05 Thread Johnny Tan
# cat /sys/kernel/mm/redhat_transparent_hugepage/defrag [always] never On Tue, Feb 5, 2013 at 5:37 PM, Josh Krupka wrote: > Just out of curiosity, are you using transparent huge pages? > On Feb 5, 2013 5:03 PM, "Johnny Tan" wrote: > >> Server specs: >> Dell R610 >> dual E5645 hex-core 2.4GHz >

Re: [PERFORM] postgresql.conf recommendations

2013-02-05 Thread Josh Krupka
Just out of curiosity, are you using transparent huge pages? On Feb 5, 2013 5:03 PM, "Johnny Tan" wrote: > Server specs: > Dell R610 > dual E5645 hex-core 2.4GHz > 192GB RAM > RAID 1: 2x400GB SSD (OS + WAL logs) > RAID 10: 4x400GB SSD (/var/lib/pgsql) > > > /etc/sysctl.conf: > kernel.msgmnb = 655

[PERFORM] postgresql.conf recommendations

2013-02-05 Thread Johnny Tan
Server specs: Dell R610 dual E5645 hex-core 2.4GHz 192GB RAM RAID 1: 2x400GB SSD (OS + WAL logs) RAID 10: 4x400GB SSD (/var/lib/pgsql) /etc/sysctl.conf: kernel.msgmnb = 65536 kernel.msgmax = 65536 kernel.shmmax = 68719476736 kernel.shmall = 4294967296 vm.overcommit_memory = 0 vm.swappiness = 0 vm

Re: [PERFORM] Slow Query Help

2013-02-05 Thread Heikki Linnakangas
On 05.02.2013 05:45, Will Platnick wrote: We upgraded from PG 9.1 to 9.2. Since the upgrade, the # of active queries has raised significantly, especially during our peak time where lots of users are logging in. According to New Relic, this query is now taking up the most amount of time during