Large dirs [Was: Re: Clustering and replication]

[Tom Samplonius]

----- "Janne Peltonen" <janne.peltonen at helsinki.fi> wrote:
> On Tue, Jan 30, 2007 at 11:01:03AM +0100, Simon Matter wrote:
> > > This might be a problem, since we have some users that really have
> 10000
> > > messages in their INBOX. Although it seems that Cyrus itself
> cannot cope
> > > with this either... in our current, non-clustering setup. But
> then, it's
> > > an old version.
> > Interesting, I have some folders with ~100'000 messages and cyrus
> handles
> > it very nice. Did you say you have a problem with 10'000 messages in
> a
> > mailbox?
> 
> I didn't inspect the case at all closely, but yes, there was one user
> with some 10'000 messages in one folder experiencing slowdown, and
> after
> I suggested him to clean up his INBOX, there were no more complaints.
> But it was (is) really an old version of Cyrus on a very busy server.

  I've got 17,500 e-mail messages in my Inbox on a Cyrus 1.5 system (that is old), and have not seen this problem.

  It could be filesystem and kernel related.  ext2/3 has historically not been good at this, but is better now.  ufs is actually pretty good, but mainly because BSD kernels cache so much vnode data, it does not matter if the directory is hashed on disk or not.  

  And you always want "noatime" enabled on your Cyrus spool, no matter the filesystem or OS, or version of Cyrus.

Tom