- Previous message: Peter A. Friend: "Re: Cyrus Quota vs. du"
- Maybe in reply to: Dana Canfield: "Cyrus Quota vs. du"
- Next in thread: Simon Matter: "Re: Cyrus Quota vs. du"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
|
|
|
From: Andrew Morgan (no email)
Date: Wed May 02 2007 - 12:48:20 EDT
On Wed, 2 May 2007, Peter A. Friend wrote:
> On May 2, 2007, at 8:51 AM, Dana Canfield wrote:
>
>> We've been slowly phasing in mail quotas and I've discovered that the quota
>> usage reported by cyrus doesn't match a du for that same user. In fact, I
>> can't find any user where the reports *do* match. Is this a sign of quota
>> corruption, or am I overlooking something about how the quotas work?
>>
>> Here are some examples to show what I'm checking:
>>
>> [root at balrog ~]# du -s /var/spool/imap/e/user/efreem
>> 1068664 /var/spool/imap/e/user/efreem
>> [root at balrog ~]# du -s /var/spool/imap/a/user/austin
>> 1178924 /var/spool/imap/a/user/austin
>>
>> [root at balrog efreem]# cyradm --user cyrus localhost
>> localhost.localdomain> lq user.efreem
>> STORAGE 904193/2000000 (45.20965%)
>> localhost.localdomain> lq user.austin
>> STORAGE 703056/1000000 (70.3056%)
>
> The du utility reports usage in terms of blocks, not bytes. This is a major
> drawback of maildir type systems, they can have nasty partial block
> overheads.
Also, 'du' will include the size of the cyrus.index, cyrus.header, and
cyrus.cache files. Cyrus' quota does not include those files.
Andy
---- Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
|
|
|