Horton

Re: [HACKERS] Performance while loading data and indexing

Bruce Momjian <pgman@candle.pha.pa.us>

View thread View on postgr.es
From: Bruce Momjian <pgman@candle.pha.pa.us>
To: Rod Taylor <rbt@rbt.ca>
Cc: Neil Conway <neilc@samurai.com>, Greg Copeland <greg@CopelandConsulting.Net>, shridhar_daithankar@persistent.co.in, PostgresSQL Hackers Mailing List <pgsql-hackers@postgresql.org>, PostgresSQL General Mailing List <pgsql-general@postgresql.org>
Date: 2002-09-26T22:04:52Z
Lists: pgsql-hackers, pgsql-performance, pgsql-general 
Rod Taylor wrote:
> On Thu, 2002-09-26 at 17:47, Bruce Momjian wrote:
> > Rod Taylor wrote:
> > > > Yes, before UFS had soft updates, the synchronous nature of UFS made it
> > > > slower than ext2, but now with soft updates, that performance difference
> > > > is gone so you have two files systems, ext2 and ufs, similar peformance,
> > > > but one is crash-safe and the other is not.
> > > 
> > > Note entirely true.  ufs is both crash-safe and quick-rebootable.  You
> > > do need to fsck at some point, but not prior to mounting it.  Any
> > > corrupt blocks are empty, and are easy to avoid.
> > 
> > I am assuming you need to mount the drive as part of the reboot.  Of
> > course you can boot fast with any file system if you don't have to mount
> > it.  :-)
> 
> Sorry, poor explanation.
> 
> Background fsck (when implemented) would operate on a currently mounted
> (and active) file system.  The only reason fsck is required prior to
> reboot now is because no-one had done the work.
> 
> http://www.freebsd.org/cgi/man.cgi?query=fsck&sektion=8&manpath=FreeBSD+5.0-current
> 
> See the first paragraph of the above.

Oh, yes, I have heard of that missing feature.

-- 
  Bruce Momjian                        |  http://candle.pha.pa.us
  pgman@candle.pha.pa.us               |  (610) 359-1001
  +  If your life is a hard drive,     |  13 Roberts Road
  +  Christ can be your backup.        |  Newtown Square, Pennsylvania 19073