| From: | Steve Horn <steve(at)stevehorn(dot)cc> |
|---|---|
| To: | pgsql-novice(at)postgresql(dot)org |
| Cc: | mclark <mclark(at)n-focus(dot)com>, Greg Cordle <gcordle(at)n-focus(dot)com> |
| Subject: | Large update and disk usage |
| Date: | 2012-04-13 14:57:00 |
| Message-ID: | CAFLkBaXTcF=vi=MoCKf1_qn03QRjwCjvP1GBc_rX6wd2zOBCdg@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-novice |
(Postgres 9.1 on CentOS)
Performing an update to two columns on a table with 40 million records, all
in one transaction.
The size of the table on disk (according to pg_relation_size) is 131GB. My
question is: when an update to all of these rows is performed, how much
disk space should I provision?
Also would be nice to understand how Postgres physically handles large
updates like this. (Does it create a temporary or global temporary table,
and then drop it when the transaction is committed?)
--
Steve Horn
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Andreas Kretschmer | 2012-04-13 15:35:48 | Re: Large update and disk usage |
| Previous Message | Stephen Cook | 2012-04-13 11:20:58 | Re: Service start up error "The Service name is invalid net helpmsg 2185" |