Re: [PERFORM] Massive I/O spikes during checkpoint

2012-07-10 Thread Andres Freund
On Tuesday, July 10, 2012 03:36:35 PM Jeff Janes wrote: > On Tue, Jul 10, 2012 at 5:44 AM, Andres Freund wrote: > > On Tuesday, July 10, 2012 08:14:00 AM Maxim Boguk wrote: > >> So kernel doesn't start write any pages out in background before it has > >> at least 13Gb dirty pages in kernel memory

Re: [PERFORM] Massive I/O spikes during checkpoint

2012-07-10 Thread Jeff Janes
On Tue, Jul 10, 2012 at 5:44 AM, Andres Freund wrote: > On Tuesday, July 10, 2012 08:14:00 AM Maxim Boguk wrote: >> >> So kernel doesn't start write any pages out in background before it has at >> least 13Gb dirty pages in kernel memory. >> And at end of the checkpoint kernel trying flush all dirt

Re: [PERFORM] Massive I/O spikes during checkpoint

2012-07-10 Thread Andres Freund
On Tuesday, July 10, 2012 08:14:00 AM Maxim Boguk wrote: > On Tue, Jul 10, 2012 at 4:03 PM, David Kerr wrote: > > On Jul 9, 2012, at 10:51 PM, Maxim Boguk wrote: > >> But what appears to be happening is that all of the data is being > >> written out at the end of the checkpoint. > >> > >> This ha

Re: [PERFORM] Massive I/O spikes during checkpoint

2012-07-09 Thread David Kerr
On 7/9/2012 11:14 PM, Maxim Boguk wrote: On Tue, Jul 10, 2012 at 4:03 PM, David Kerr mailto:d...@mr-paradox.net>> wrote: On Jul 9, 2012, at 10:51 PM, Maxim Boguk wrote: But what appears to be happening is that all of the data is being written out at the end of the chec

Re: [PERFORM] Massive I/O spikes during checkpoint

2012-07-09 Thread Maxim Boguk
On Tue, Jul 10, 2012 at 4:03 PM, David Kerr wrote: > > On Jul 9, 2012, at 10:51 PM, Maxim Boguk wrote: > > >> >> But what appears to be happening is that all of the data is being written >> out at the end of the checkpoint. >> >> This happens at every checkpoint while the system is under load. >>

Re: [PERFORM] Massive I/O spikes during checkpoint

2012-07-09 Thread David Kerr
On Jul 9, 2012, at 10:51 PM, Maxim Boguk wrote: > > > But what appears to be happening is that all of the data is being written out > at the end of the checkpoint. > > This happens at every checkpoint while the system is under load. > > I get the feeling that this isn't the correct behavior

Re: [PERFORM] Massive I/O spikes during checkpoint

2012-07-09 Thread David Kerr
On Jul 9, 2012, at 10:52 PM, Jeff Janes wrote: > On Mon, Jul 9, 2012 at 10:39 PM, David Kerr wrote: >> >> I thought that the idea of checkpoint_completion_target was that we try to >> finish writing >> out the data throughout the entire checkpoint (leaving some room to spare, >> in my case 30%

Re: [PERFORM] Massive I/O spikes during checkpoint

2012-07-09 Thread Jeff Janes
On Mon, Jul 9, 2012 at 10:39 PM, David Kerr wrote: > > I thought that the idea of checkpoint_completion_target was that we try to > finish writing > out the data throughout the entire checkpoint (leaving some room to spare, > in my case 30% > of the total estimated checkpoint time) > > But what ap

Re: [PERFORM] Massive I/O spikes during checkpoint

2012-07-09 Thread Maxim Boguk
> > > > But what appears to be happening is that all of the data is being written > out at the end of the checkpoint. > > This happens at every checkpoint while the system is under load. > > I get the feeling that this isn't the correct behavior and i've done > something wrong. > > > It's not an ac

[PERFORM] Massive I/O spikes during checkpoint

2012-07-09 Thread David Kerr
Howdy! I'm trying to figure out why checkpointing it completely pegging my I/O under moderate to high write load, I'm on PG9.1.1, RHEL 6.2 x64 checkpoint_completion_target = 0.7 checkpoint_timeout = 10m Jul 10 00:32:30 perf01 postgres[52619]: [1895-1] user=,db= LOG: checkpoint starting: time