I've got about 50K boxes on NFS with maildir++. I'm running vpopmail 5.2.1 though. The only problem I've ever seen is every once in a great while Courier-IMAP will return a wildly wrong negative value for number of unread messages, which breaks some other stuff I run, but I doubt there's a correlation.

Nick Harring

Tim Hasson wrote:

This is a followup for the problem discussed last week.

Is there anyone even running maildir on nfs with maildir++ quotas enabled??

----- Forwarded message from Sam Varshavchik <[EMAIL PROTECTED]> ----- Date: Mon, 22 Sep 2003 23:55:26 -0400 From: Sam Varshavchik <[EMAIL PROTECTED]> Reply-To: Sam Varshavchik <[EMAIL PROTECTED]> Subject: Re: NFS, maildir++ and -ve quota To: Tim Hasson <[EMAIL PROTECTED]>

Tim Hasson writes:

Quoting Sam Varshavchik <[EMAIL PROTECTED]>:

Tim Hasson writes:

Mr Sam,

I have switched over to maildrop instead of vdelivermail, by calling it from .qmail-default for all my domains. Everything looked good for a few


until I see the problem occuring with one of the users that had the same problem before with vdelivermail.

So I ran a shell script that removes any maildirsize file with negatives


it, then recreate the maildirsize file for that user by running "vuserinfo


[EMAIL PROTECTED]" and setting the chown/chgrp vpopmail/vchkpw on the


file. Just in case some bad stuff stayed around..

I am not familiar with vuserinfo. I can tell you, though, that maildirmake -q should recalculate the current quota.

The `vuserinfo -Q [EMAIL PROTECTED] calculates the quota, displays usage percentage on screen, and creates maildirsize in the user's maildir if it


not already exist.

Few days, and the problem happens to another account. There is 1.1MB in user/Maildir, his trash is only few kb, but getquota ROOT on courierimap reports -ve value.

What exactly is a "-ve" value?

Lousy abbreviations: (-)ve == negative

Well, somehow messages that are added to the maildir are not added to the quota, so when the messages are removed, the subtracted quota becomes negative.

I don't have an immediate idea why. Although Maildir++ certainly trades in some known race conditions in favor of lock-free operation, race condition that would lead to such a situations should be very rare. Perhaps NFS exaggerates the conditions that lead to a race condition.

But that's pure speculation. What's needed is a reproducible way to create this scenario.

One potential workaround you can explore is to forcibly recompute the quota if the apparent quota is negative.

Try hacking the quota code to call maildir_quota_recalculate(), just once,
if the apparent quota is negative, then try again.

----- End forwarded message -----

Respectfully, Tim Hasson

Reply via email to