On 2006-10-13, Alexander Staubo <alex@purefiction.net> wrote:
> On my box (Dell PowerEdge 1850, dual Xeon 2.8GHz, 4GB RAM, 10kRPM
> SCSI, Linux 2.6.15, Ubuntu) I get 1,100 updates/sec, compared to
> 10,000 updates/sec with MySQL/InnoDB, using a stock installation of
> both. Insert performance is only around 10% worse than MySQL at
> around 9,000 rows/sec. Curiously enough, changing shared_buffers,
> wal_buffers, effective_cache_size and even fsync seems to have no
> effect on update performance, while fsync has a decent effect on
> insert performance.
Your disk probably has write caching enabled. A 10krpm disk should be
limiting you to under 170 transactions/sec with a single connection
and fsync enabled.
I also did some tests on this, and even though the machine I was testing
on had some competing database activity, autovacuum was effective at
keeping the table size stable (at 70-odd pages) when running several
hundred thousand updates on a 1-row table.
--
Andrew, Supernews
http://www.supernews.com - individual and corporate NNTP services