2010/7/2 Timo Sirainen <t...@iki.fi>: > On Fri, 2010-07-02 at 14:31 +0200, Joan wrote: >> > Did you try the quota recalculation when logging in normally, instead of >> > with --exec-mail? >> I sent a 5M mail to myself, there's the mail in inbox and the copy to >> sent (amounting 10M), I connected to the server and manually removed >> the mail (rm file). >> After that sending a new mail updates the quota, but it doesn't recalculate >> it. > > Right. Quota isn't automatically recalculated if you go manually adding > or removing files. I see, I expected to have something like when I delete the maildirsize file (when using maildir++ quotas) and dovecot automatically recalculates the quota. > >> After deleting the entry for t...@testing.com from the database, the >> quota gets properly recalculated. > > So the problem exists only with --exec-mail? Yes, you're right, the proper quotas don't get calculated with exec-mail and most of the users won't enter ever via webmail and only download mails via pop > >> The problem is that I rely on db quota for some accounting scritpts, >> and the users might not enter to their webmail or download the mail >> for days. >> >> How could I trigger this quota recalculation? > > If the problem is only with --exec-mail, you could create a master user > that can login to any user's account and triggers the quota > recalculation. Something like: > > printf "a login destuser*masteruser masterpass\nb getquotaroot inbox\n" | nc > localhost 143 > > http://wiki.dovecot.org/Authentication/MasterUsers Wow, yet another amazing feature of dovecot. This is really a swiss knife mailserver :P I will try to purge the wrong quotas and see what happens when new mail is delivered (it's better no quota than wrong quota) > >> .- Maybe the most practical would be to have a crontab to update the >> real values from time to time, at least until 2.0 is not available in >> our environment... > > In v2.0 you can do it even more easily :) > doveadm quota recalc -u username I hope I can get there soon :)
Thanks again
