I've just found that in at least one user's Maildir some of the files in the cur directory appear 'flashing'(!) in red in my terminal and the same files have their properties/attributes replaced with question marks.

I discovered them because I ran du -hs on the home directory and those files were listed as inaccessible even to root.

Here is the listing:

drwxrwx--- 2 travel mail 8.0K Aug 11 19:29 .
drwxrwx--- 5 travel mail 4.0K Aug 11 19:30 ..
-rw-rw---- 1 travel mail  33K Jun 27 03:55 1327311651.M313310P1633.my.domain.info,W=2330:2,STa
?--------- ? ?         ?       ?            ? 1327378561.M795425P7819.my.domain.info,W=4603093:2,ST
?--------- ? ?         ?       ?            ? 1327408110.M597988P17812.my.domain.info,W=3910:2,ST
?--------- ? ?         ?       ?            ? 1327665200.M938383P29773.my.domain.info,W=5762:2,ST
?--------- ? ?         ?       ?            ? 1327904625.M697896P18082.my.domain.info,W=7674:2,ST

The machine is a VPS running CentOS 5.


The output of mount:

# mount
/dev/vzfs on / type reiserfs (rw,usrquota,grpquota)
proc on /proc type proc (rw,relatime)
sysfs on /sys type sysfs (rw,relatime)
none on /dev type tmpfs (rw,relatime)
none on /dev/pts type devpts (rw,relatime)
none on /proc/sys/fs/binfmt_misc type binfmt_misc (rw,relatime)

The contents of fstab:

# cat /etc/fstab
# none  /dev/pts        devpts  rw,usrquota,grpquota    1       1
none    /dev/pts        devpts  rw      0       0

The result of du -h:

# du -h /home/cust243/imap/custdomain.com.ph/info/Maildir/.INBOX.Sent/
4.0K    /home/cust243/imap/custdomain.com.ph/info/Maildir/.INBOX.Sent/tmp
4.0K    /home/cust243/imap/custdomain.com.ph/info/Maildir/.INBOX.Sent/new
du: cannot access `/home/cust243/imap/custdomain.com.ph/info/Maildir/.INBOX.Sent/cur/1327665200.M938383P29773.s1.ourdomain.info,W=5762:2,ST': No such file or directory
du: cannot access `/home/cust243/imap/custdomain.com.ph/info/Maildir/.INBOX.Sent

### Truncated many more similar files ###

du: cannot access `/home/cust243/imap/custdomain.com.ph/info/Maildir/.INBOX.Sent/cur/1327378561.M795425P7819.s1.ourdomain.info,W=4603093:2,ST': No such file or directory
25M     /home/cust243/imap/custdomain.com.ph/info/Maildir/.INBOX.Sent/cur
25M     /home/cust243/imap/custdomain.com.ph/info/Maildir/.INBOX.Sent/

This is preventing the customer from being able to list their sent emails, so I need to clean it up somehow but not sure how since this is what happens:

# rm /home/cust243/imap/custdomain.com.ph/info/Maildir/.INBOX.Sent/cur/1327665200.M938383P29773.s1.ourdomain.info,W=5762:2,ST
rm: cannot lstat `/home/cust243/imap/custdomain.com.ph/info/Maildir/.INBOX.Sent/cur/1327665200.M938383P29773.s1.ourdomain.info,W=5762:2,ST': No such file or directory

There are also a bunch of files in the directory which appear to be un-effected by the corruption.

  • 1
    Can you post the output of mount as well as cat /etc/fstab and df -h? This looks at first glance like an underlying filesystem issue - the only times I've seen this behaviour is on a broken NFS or SMBFS/CIFS mount. – Craig Watson Aug 11 '13 at 14:15
  • @CraigWatson I've added the additional info to my question. IS this something that can be fixed from within the container, I'm wondering, or do I need to contact the VPS provider to resolve it. It does look to me, like it might have been caused by a physical disk problem at some stage. – Peter Snow Aug 11 '13 at 14:45

As I said in my comment, this looks to be an issue with the underlying filesystem within your VPS.

There are no network filesystems mounted, so those aren't to blame. The next step would be to run an offline fsck to rule out the possibility of your VPS' filesystem becoming corrupt.

As your server is a VPS, you will need the ability to access the VPS' console, or at least boot the server into a rescue system capable of running an fsck.

To do this, your VPS will need to be shut down, so you should warn your customers that service will be interrupted - depending on the size and underlying hardware of the disk, this could take multiple hours.

Either way, you should contact your VPS provider and make them aware - they may have had issues with their infrastructure that has caused the filesystem corruption.

  • Ok. I kind of suspected that, but wanted a second opinion in case I'd missed something simple. I plan to build a replacement (it's due anyway) and switch over to it without the users noticing... That's plan anyway! – Peter Snow Aug 11 '13 at 16:47

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.