| From: | Brian Ghidinelli <brian(at)pukkasoft(dot)com> |
|---|---|
| To: | SF Postgres <sfpug(at)postgresql(dot)org> |
| Subject: | Starting point for upsizing 8.3 -> 9.0 and 6GB -> 16GB |
| Date: | 2011-01-12 19:56:58 |
| Message-ID: | 4D2E078A.2080108@pukkasoft.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | sfpug |
Hi list, we're making a big upgrade from our in-house 8.3.7 server to a
managed 9.0.2 server. The hardware is from 6GB to 16GB of RAM. Would
anyone care to make some guesstimates for upsizing our config? We're
going to test but it would be nice to start in the ballpark.
Today on 6GB RAM:
shared_buffers = 1024MB
work_mem = 4MB
maintenance_work_mem = 256MB
wal_buffers = 1MB
effective_cache_size = 3000MB (no longer exists in 9?)
checkpoint_segments = 10
My thought is to increase shared_buffers to 6144MB, remove
effective_cache_size and leave the rest?
Also, we've enabled hot standby on the new 9 server (but don't yet have
a slave). I noticed that there are ~230 16mb files in the archive
directory. checkpoint_segments on that box is not set so I would think
it would be some (much smaller) multiple of 3, no? The hot standby
directory is basically the same size as the data dir (~5GB).
Many thanks!
Brian
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Josh Berkus | 2011-01-21 05:37:49 | Re: Starting point for upsizing 8.3 -> 9.0 and 6GB -> 16GB |
| Previous Message | Brian Ghidinelli | 2010-12-30 07:05:44 | Automating Hot Standby |