Re: Inserting 8MB bytea: just 25% of disk perf used?

On 19/01/10 10:50, fkater(at)googlemail(dot)com wrote:
> However, the deeper question is (sounds ridiculous): Why am
> I I/O bound *this much* here. To recall: The write
> performance in pg is about 20-25% of the worst case serial
> write performance of the disk (and only about 8-10% of the
> best disk perf) even though pg_xlog (WAL) is moved to
> another disk, only 10 simple INSERT commands, a simple table
> of 5 columns (4 unused, one bytea) and one index for OID, no
> compression since STORAGE EXTERNAL, ntfs tweaks (noatime
> etc), ...

I'm no Windows expert, but the sysinternals tools (since bought by
Microsoft) have always proved useful to me.

Diskmon should show you what's happening on your machine:
http://technet.microsoft.com/en-us/sysinternals/bb896646.aspx

Be aware that this will generate a *lot* of data very quickly and you'll
need to spend a little time analysing it. Try it without PG running to
see what your system is up to when "idle" first to get a baseline.

Unfortunately it doesn't show disk seek times (which is probably what
you want to measure) but it should let you decode what reads/writes are
taking place when. If two consecutive disk accesses aren't adjacent then
that implies a seek of course. Worst case you get two or more processes
each accessing different parts of the disk in an interleaved arrangement.

--
Richard Huxton
Archonet Ltd