I have the flowing config after adding a new drive and moving web/backup to that disk. Code: # /etc/fstab: static file system information. proc /proc proc nodev,noexec,nosuid 0 0 /dev/vda1 / ext3 errors=remount-ro,usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0 0 1 /dev/vdb /diskb/ ext4 errors=remount-ro,usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0 0 2 /diskb/www /var/www none bind,nobootwait,_netdev,usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0 0 0 /diskb/backup /var/backup none bind,nobootwait,_netdev 0 0 Everything worked fine for a couple of months until recently. Now some customers sites soft and hard limits show up doubled in the ISPConfig dashboard. Seams like everyone I change their space on a certain web gets doubled so, everything new get their soft and hard limit doubled. But! Even though it seams doubled in the dashboard/visual meter the quota still works on it's actual level. For example I now had a customer with 50000MB set as soft limit and 50001MB set as hard limit. This night they got a quota warning since they've hade moved past their set 50GB. In the dashboard it's still just yellow since their soft limit shows 100000MB and hard limit shows 100002MB. Any ideas on how to fix this?
*Bump* No on who knows why this is happening? Seams like the quotas besides the visual dashboard works as they should. A customer with 5 000 quota who shows up as 10 000 in the visual dashboard got a warning now when they came close to 5 000.
Linux filesystem quota works per hard disk, it's not an overall quota. When you have two harddisks or more and enable quota on them, then the user has x amount of quota on each of them. The dashboard shows his overall quota, which is 10 000 while he has 5 000 on each harddisk and what you see as !quota works" means that he reached 5 000 on one harddisk and was denied to add more. If you would have tried to add data for this user on the second harddisk, then you would have noticed that this is possible. So the ISPConfig dashboard is correct, it shows the total amount of space that the user might use on the server, and the Linux quota system is correct as well, as it denied to add more data on one harddiks while the user would be able to add more data on the second one. The solution is to remove linux quota in fstab from the harddisk that is not used for the websites (/var/www).
Tanks, that makes sense! But we still have email under /dev/vda1 that we want quota on. Then again, you wrote /var/www and that got me a bit confused. How would a correct fstab look, keeping the quota on emails on the main disk and at the same time have a correct quota on the web on diskb?
Email does not need Linux quota. The email quota is enforced by the IMAP server (courier or dovecot). So you can remove quota from mail partition without affecting email quota.
Hey @till How do I modify it so I don't wreak something else? Code: # /etc/fstab: static file system information. proc /proc proc nodev,noexec,nosuid 0 0 /dev/vda1 / ext3 errors=remount-ro,usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0 0 1 /dev/vdb /diskb/ ext4 errors=remount-ro,usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0 0 2 /diskb/www /var/www none bind,nobootwait,_netdev,usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0 0 0 /diskb/backup /var/backup none bind,nobootwait,_netdev 0 0 Remove "usrjquota=aquota.user,grpjquota=aquota.group" on /dev/vda1 / ext3 ? Or is it the double one for /dev/vdb /diskb/ ext4 + /diskb/www /var/www ?
@till wrote Email does not need Linux quota so you can remove the quota from the partition that holds mailboxes (or Maildirs). Figure out which partion that is and remove the quota stuff from that line in the /etc/fstab. Make a copy of the /etc/fstab file first, so you can boot in single user mode or from Live CD if the mess prevents normal booting. I take no responsibility of your server, I don't know what hardware or what operating system you are running so it is all your responsibility if you follow random advice from the Internet and a great catastrophe appears.
Basically, you should be able to use the steps from perfect server, just in reverse order and by using off instead of on: 1) Turn quota off: quotaoff -avug 2) Remove the quota part from the / partition that contains the mail data in fstab. 3) remount the partition: mount -o remount / 4) Turon on quota again qith: quotaon -avug Untested and no guarantee
New fstab: Code: # /etc/fstab: static file system information. proc /proc proc nodev,noexec,nosuid 0 0 /dev/vda1 / ext3 errors=remount-ro 0 1 /dev/vdb /diskb/ ext4 errors=remount-ro,usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0 0 2 /diskb/www /var/www none bind,nobootwait,_netdev,usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0 0 0 /diskb/backup /var/backup none bind,nobootwait,_netdev 0 0 Seams to have done the trick. Thanks!
