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

Top Page
Delete this message
Reply to this message
Author: Andreas Metzler
Date:  
To: 'exim-users@exim.org'
Subject: Re: [Exim] Large queue sizes - big directories - and Reiser File System
On Thu, Dec 04, 2003 at 09:38:16AM -0800, Marc Perkel wrote:
> Andreas Metzler wrote:
> >On Thu, Dec 04, 2003 at 08:11:43AM -0800, Marc Perkel wrote:
>>> For those of you who are suffering for slow server when you get 100,000
>>> + messages in the queue and want to know what you can do to speed up
>>> your server - dump ext2/ext3 and go with the reiser file system.
>> [...]
>> Or stay with ext3 and enable dir_index/htree.


>> dir_index is present in 2.6.* and for example in the (patched) Kernels
>> recent versions of RedHat ship.


> What do you have to do to enable that


1. Run a kernel and e2fsutils that support it.
2. switch it on with tune2fs
3. Iirc there is some kind of conversion for existing big (>= 1024
entries) directories needed, otherwise they stay slow. However I
can't find the instructions right now. - 'cp -a orig copy;
mv orig old; mv copy orig' and removing old afterwards should
do the trick.

> and how well does it work?


It uses hashed b-trees (like ReiserFS) and is therfore rather fast.
               cu andreas