Re: [Exim] Re: Large queue sizes - big directories - and Rei…

Top Page
Delete this message
Reply to this message
Author: Nigel Metheringham
Date:  
To: Marc Perkel
CC: exim-users
Subject: Re: [Exim] Re: Large queue sizes - big directories - and Reiser File System
On Fri, 2003-12-05 at 19:31, Marc Perkel wrote:
> I've run Reiser for 3 years on several servers and never had a problem
> with it.


This has actually been one of the major problems with reiser from the
very start. There have always been reports of the type "I'm using
reiser and it rocks" and these have been used as evidence of
reliability. What always appears to have been lacking is evidence of
resilience and recoverability when something goes wrong.

Ext3 is highly recoverable - I have only once seen a filesystem trashed
to the point where I decided it was easier to use mkfs as a recovery
tool and in that case I knew I was running a very early alpha of a
particular feature set. I have heard (yes anecdotal - but one of the
sites had been a serious reiser proponent for ages) cases where people
have just lost reiser partitions for no apparent reason, and were
completely unable to recover without a reformat and restore.

I am not trying to diss reiser - I have just built a server with reiser
filesystems as an experiment. However I have consistently had the
feeling that they put features and speed first, reliability and
recoverability (which includes from minor hardware glitches) a distant
second.

    Nigel.


--
[ Nigel Metheringham           Nigel.Metheringham@??? ]
[ - Comments in this message are my own and not ITO opinion/policy - ]