OSEC

Neohapsis is currently accepting applications for employment. For more information, please visit our website www.neohapsis.com or email hr@neohapsis.com
 
Subject: RE: large postfix installations?
From: Thomas Reagan (tkrbrown.edu)
Date: Tue Apr 04 2000 - 15:16:48 CDT

That's why most controllers that enable this feature (read: RAID
controllers) have a battery-backed cache which will flush the next time
power is restored to the drives.

-----Original Message-----
From: owner-postfix-userspostfix.org
[mailto:owner-postfix-userspostfix.org]On Behalf Of Ian C.Sison
Sent: Tuesday, April 04, 2000 3:08 PM
To: Postfix Users List
Subject: Re: large postfix installations?

On Tue, 04 Apr 2000, you wrote:
> Ian C.Sison <ian.sqsr.com.ph> writes:
>
> > > Better yet, switch to FreeBSD and use softupdates, where you
> > > should see a factor of five to ten performance improvement.
> >
> > Question on this. Isn't the issue on async updates vs sync updates,
> > on any OS moot if your disk controller or HD contains a large enough
> > cache that buffers writes to the disk anyway?
>
> Not really. With classical systems, sync update means that the whole lot
> you're writing has been actually written to the physical disk until your
> write or close or fsync calls return (depending which way you enforce
> data synchronization).

I'm probably missing specific behaviour here, but if a controller or hard
disk caches writes, and fools the OS into believing that the data the OS
asked
the controller to write has been written to the disk [which is most probably
the
case whether the OS is in sync or async write mode], and power goes out, the
same data corruption and integrity problems will exist.