I've a problemc oncerning "quota". When a user is over quota, he can no more access his mailbox. We are using Maildir and webimiau access mail server via pop3 ... Also IMAP client (es: thunderbird) blocks! (and so squirrelmail too) Is there a solution to prevent this? Another answer: in our setting uebimiau saves users mails in /home/admispconfig/ispconfig/temp/webmail where files belongs to admispconfig:admispconfig So mail is not considered in user quota, is it? Can I configure webimiau to use Imap? In this way users's mails remains all in Maildir??? Now we have both squirrelmail (using imap) and uebimiau (using pop3) and folders created by the two programs are not seen by the other! And mail moved in a folder of one are no more visible from the other. If we pass uebimiau to "imap" this problem is resolved? thanks!!! Giorgio
Are you using the mbox or the maildir format? afair the problem with mbox is if you donwload your mails you need the double space currently used by the mails in you mailbox.... e.g. mails of 60MB, quota till 100, upon downloading them you will be 20Megs over quota...
This happens normally only with mailbox, not maildir as Ben posted. These copies of the emails are not in quota. no. Use a maildir capable webmail client like squirrelmail. No.
>Originally Posted by giorgio.volpe >>I've a problem concerning "quota". >> >>When a user is over quota, he can no more access his mailbox. >> >>We are using Maildir and webimiau access mail server via pop3 ... >>Also IMAP client (es: thunderbird) blocks! (and so squirrelmail too) >> >>Is there a solution to prevent this? > >This happens normally only with mailbox, not maildir as Ben posted. We do use Maildir! When user is over quota imap doesn't work any more! May be this is a problem of postfix ( in fact also external client as thunderbird blocks).
Postfix doesn't have anything to do with IMAP so it's not a Postfix problem. What's the output of Code: netstat -tap ? To solve your current problem you could increase the user's quota, let him fetch his mails and set the quota back to its original value.
Ok ... but we have many user and that it's a problem! (furthermore IspConfig is very slow to write config files in a web with near 3000 users, change are really heavy)
I apologise ... I send again my replay with correct "quoting" You are right: it's courier imap that doesn't work when user is over quota. I imagine it needs to create some lock or cache in user dierectory. I wonder if it's possible to tell postfix to reject new mail when user is near his quota ... so that he can still use imap to delete mails and free some space. What's the output of Code: netstat -tap ? Very long ... You need this? tcp6 0 0 *:imaps *:* LISTEN 7438/couriertcpd tcp6 0 0 *:imap2 *:* LISTEN 15234/couriertcpd Ok ... but we have many user and that it's a problem! (furthermore IspConfig is very slow to write config files in a web with near 3000 users, change are really heavy)
Ok, you're using Maildir. Did you enable Maildir in ISPConfig (Management -> Server -> Settings -> Email)?
I cut only an extract from the 11941 row of the output ... *** Report for user quotas on device /dev/mapper/vg1-home Block grace time: 7days; Inode grace time: 7days Block limits File limits User used soft hard grace used soft hard grace ---------------------------------------------------------------------- root -- 9406820 0 0 88297 0 0 mail -- 4 0 0 1 0 0 www-data -- 735528 0 0 64839 0 0 admispconfig -- 1487040 0 0 42180 0 0 .... web80_posta -- 80 5120 5120 14 0 0 web67_info -- 452 0 0 14 0 0 web67_sve -- 28 0 0 8 0 0 ... web43_apasqualetti +- 5228 5120 5120 6days 70 0 0 ... #22185 -- 36 0 0 9 0 0 #22678 -- 996 0 0 9 0 0 #23171 -- 136 0 0 14 0 0 ... Statistics: Total blocks: 21 Data blocks: 11 Entries: 225 Used average: 20,454545 *** Report for group quotas on device /dev/mapper/vg1-var Block grace time: 7days; Inode grace time: 7days Block limits File limits Group used soft hard grace used soft hard grace ---------------------------------------------------------------------- root -- 5850500 0 0 3773 0 0 mail -- 1144 0 0 2112 0 0 www-data -- 18564 0 0 3579 0 0 ... web9 -- 0 1024000 1024000 0 0 0 web10 -- 0 614400 614400 0 0 0 web11 -- 0 51200 51200 0 0 0 ... Statistics: Total blocks: 13 Data blocks: 6 Entries: 121 Used average: 20,166667 PS: what are rows like #12345 ?
That user is over quota, therefore he cannot fetch his mails... 12345 is the user id of a user that doesn't exist anymore in /etc/passwd. The user was deleted, but some files/directories are still owned by him.
Ok ... let's make a summary of the problem: 1) if a user is over quota He can no more use webmail (nor an imap client) even to delete some mail and so the only way should be to fetch and delete some mail via pop3. (or, worse, if also pop3 blocks to change temporarly user quota ...) 2) This is not a problem of Ispconfig or Uebimiau or Squirrelmail ... rather it's a problem of courier Imap that needs some space to operate over Maildir data creating at least some lock file or other temporary store for its work. I've seen in each user directory a file called .quota.rc... We could resolve the problem modifing this file and refusing mails when user space is used, for exampel, at 95%? So new mail will no more block courier ... Further more we could reserve a little space to send an email to the user to tell him "Quota reached ... we are refusing mails! You must free some space!" Is this a good idea? If so I can try to implement it Giorgio
NB: I've seen that .quota.rc now doesn't work in our environment cause the command "quota" reports two devices and not only one as the script expects .. Giorgio # quota web43_giorgio Disk quotas for user web43_giorgio (uid 10197): Filesystem blocks quota limit grace files quota limit grace /dev/mapper/vg1-home 824 51200 51200 67 0 0 /dev/mapper/vg1-var 0 51200 51200 1 0 0 So we have: # quota web43_giorgio | sed -n 3p /dev/mapper/vg1-home
I've just changed the way ISPConfig configures quota. Up to now, the quota soft limit and hard limit were identical, but in the next release ISPconfig will make the hard limit = soft limit + 1 MB, so there's a little bit space left to send the user a notification that he needs to free up space, etc.
