Zimbra Hard Disk Full

[gnyce]

Based on the below number, it looks like you do actually have some free space on /dev/md0, but is reserved (betting this is ext3 filesystem). So you couldrun tune2fs and drop-down the reserved file-space (default is 5% I think). That would give you a little breathing room, but you would of course still have to cut back on your disk usage. Some something like "tune2fs -m 2 /dev/md0" to set it to 2%

Ext3 Filesystem Tips - ArchWiki

Once you you get a handle and cut back on usage, would recommend you set it back to 5%

I think the os and zimbra are on the same volume

mail:~# df -h
Filesystem Size Used Avail Use% Mounted on
/dev/md0 108G 102G 124M 100% /
tmpfs 506M 0 506M 0% /lib/init/rw
udev 10M 68K 10M 1% /dev
tmpfs 506M 0 506M 0% /dev/shm

[afunez2009]

Run the command as root. Switch to the root user, then run the command. It will take the command a while to complete, so be patient. The run down would look something like this:

su - <press return>

<enter root's password> <press return>

cd /

find . -maxdepth 1 -exec du -sh '{}' \;

Not sure, but you may need to add -print at the end of command.

find . -maxdepth 1 -exec du -sh '{}' \; -print

I find the large files:
One is 46GB: /opt/zimbra/log/zmmtaconfig.log
Other is 49GB: /opt/zimbra/logger/db/data/mail.err

The question is: can i delete this files directly or i have to empty the logs??

What commands i have to use for this?

Help me plese!!


Thanks For your help

[king0770]

You should be able to zero out the zmmtaconfig.log file, as zimbra run:

cd ~/log/

> zmmtaconfig.log

ls -la zmmtaconfig.log

Size should be considerably less.

As far as the mail.err file, might take a look at what's writing to the file. Just a guess but maybe one of the logger tables is corrupted (just a guess though).

You could remove the mail.err file and reinitialize the logger DB by doing the following

$ zmloggerctl stop
$ rm -rf /opt/zimbra/logger/db
$ source /opt/zimbra/bin/zmshutil
$ zmsetvars
$ /opt/zimbra/libexec/zmloggerinit ${mysql_logger_root_password}
$ zmlogswatchctl start

Or...

Since the logger service is not critical you could disable the logger service.

[afunez2009]

You should be able to zero out the zmmtaconfig.log file, as zimbra run:

cd ~/log/

> zmmtaconfig.log

ls -la zmmtaconfig.log

Size should be considerably less.

As far as the mail.err file, might take a look at what's writing to the file. Just a guess but maybe one of the logger tables is corrupted (just a guess though).

You could remove the mail.err file and reinitialize the logger DB by doing the following

$ zmloggerctl stop
$ rm -rf /opt/zimbra/logger/db
$ source /opt/zimbra/bin/zmshutil
$ zmsetvars
$ /opt/zimbra/libexec/zmloggerinit ${mysql_logger_root_password}
$ zmlogswatchctl start

Or...

Since the logger service is not critical you could disable the logger service.

I clean the zmmtaconfig.log, and then mail.err go to 84kb of size, i don't known why, but is very good

Thanks a lot for your help

[dalmate]

or you can mount an other harddisk to /opt/zimbra directory (you'll have more capability)
or you can move index + message volume to an other harddisk(your old volume still be used but all new data will be stored in new harddisk)