| From: | "Kevin Grittner" <Kevin(dot)Grittner(at)wicourts(dot)gov> |
|---|---|
| To: | "francis picabia" <fpicabia(at)gmail(dot)com>, "Brad (Toronto ON CA) Nicholson" <bnicholson(at)hp(dot)com> |
| Cc: | "pgsql-admin(at)postgresql(dot)org" <pgsql-admin(at)postgresql(dot)org> |
| Subject: | Re: Best practise for upgrade of 24GB+ database |
| Date: | 2012-01-20 21:42:09 |
| Message-ID: | 4F198B510200002500044A64@gw.wicourts.gov |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-admin |
francis picabia <fpicabia(at)gmail(dot)com> wrote:
> That's great information. 9.0 is introducing streaming
> replication, so that is another option I'll look into.
We upgrade multi-TB databases in just a couple minutes using
pg_upgrade using the hard-link option. That doesn't count
post-upgrade vacuum/analyze time, but depending on your usage you
might get away with analyzing a few tables before letting users in,
and doing the database-wide vacuum analyze while the database is in
use.
One of the other options might be better for you, but this one has
worked well for us.
-Kevin
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Brian Fehrle | 2012-01-21 00:13:08 | buffers_backend climbing during data importing, bad thing or no biggie? |
| Previous Message | Kevin Grittner | 2012-01-20 21:34:46 | Re: Meta data information on tables |