Re: import performance - Mailing list pgsql-performance

From Chris
Subject Re: import performance
Date
Msg-id 44162A10.7080205@gmail.com
Whole thread Raw
In response to Re: import performance  (Tom Lane <tgl@sss.pgh.pa.us>)
Responses Re: import performance  (David Lang <david@lang.hm>)
List pgsql-performance
Tom Lane wrote:
> Chris <dmagick@gmail.com> writes:
>
>>Tons of difference :/
>
>
> Have you checked that the I/O performance is comparable?  It seems
> possible that there's something badly misconfigured about the disks
> on your new machine.  Benchmarking with "bonnie" or some such would
> be useful; also try looking at "iostat 1" output while running the
> inserts on both machines.

I'll check out bonnie, thanks.

hdparm shows a world of difference (which I can understand) - that being
the old server is a lot slower.

hdparm -t /dev/hda
/dev/hda:
  Timing buffered disk reads:   24 MB in  3.13 seconds =   7.67 MB/sec

hdparm -T /dev/hda
/dev/hda:
  Timing cached reads:   596 MB in  2.00 seconds = 298.00 MB/sec



Newer server:
hdparm -t /dev/hda
/dev/hda:
  Timing buffered disk reads:   70 MB in  3.02 seconds =  23.15 MB/sec

hdparm -T /dev/hda
/dev/hda:
  Timing cached reads:   1512 MB in  2.00 seconds = 754.44 MB/sec

> Also, are the inserts just trivial "insert values (... some constants ...)"
> or is there more to it than that?

Straight inserts, no foreign keys, triggers etc.


The only other thing I can see is the old server is ext2:
/dev/hda4 on / type ext2 (rw,errors=remount-ro)

the new one is ext3:
/dev/hda2 on / type ext3 (rw)


If it's a server issue not a postgres issue I'll keep playing :) I
thought my config was bad but I guess not.

Thanks for all the help.

--
Postgresql & php tutorials
http://www.designmagick.com/

pgsql-performance by date:

Previous
From: Frank Wiles
Date:
Subject: Re: import performance
Next
From: David Lang
Date:
Subject: Re: import performance