[SOLVED] unexpected disk usage growth

[jwilke]

Hi,

I have just now noticed a rather large disk usage growth of 6gb within 3 hrs somewhere in /opt/zimbra/

are there any obvious possibilities for this to have happened ?

this is my current content of /opt/zimbra:

3.6M ./.saveconfig
5.3G ./db
21M ./openssl-0.9.8i
5.3M ./curl-7.18.1
822M ./logger
21M ./heimdal-1.2
12K ./contrib
3.3M ./openldap-clibs-2.3.43.7z
120M ./jdk1.5.0_17
5.0M ./tcmalloc-0.97
660K ./bin
31M ./zimbramon
25M ./sleepycat-4.2.52.6
3.9M ./cyrus-sasl-2.1.22.3z
13M ./jdk1.5.0_15
26M ./lib
3.7M ./conf
4.0K ./backup
193M ./log
1.2M ./ssl
20K ./.ssh
370M ./zmstat
724K ./docs
1.6M ./libexec
3.2M ./dspam-3.8.0
19M ./openldap-data
15M ./openldap-2.3.43.7z
22M ./postfix-2.4.7.5z
41M ./clamav-0.94.1
121M ./data
615M ./jetty-6.1.5
17M ./net-snmp-5.4.1
620K ./zimlets-experimental
1.3M ./amavisd-new-2.5.4
348K ./zimlets-network
64M ./keyview-10.5.0.0
404K ./zimlets
400K ./zimlets-extra
504K ./convertd
32M ./httpd-2.2.8
72K ./wiki
7.0G ./redolog
4.0K ./store
16G ./index
4.0K ./fbqueue
44K ./zimlets-admin-extra
4.0K ./index2
12M ./aspell-0.60.6
4.0K ./mailstore
77M ./mysql-standard-5.0.67-pc-linux-gnu-i686-glibc23
1.8M ./memcached-1.2.5
3.4M ./nginx-0.5.30-zimbra
30G .

excluding any mounted partitions (in my case /opt/zimbra/backup and /opt/zimbra/mailstore are seperate partititons)

currently /opt/zimbra is part of my root fs, and I was planning to move this to a seperate disk to prevent flooding my root fs since it was nearing 80%.

This sudden growth of another 6gb got me worried though.

[phoenix]

As you only have redolog and index folders that are greater than 6GB (that I can see there) then I'd assume there's something in there that's growing, have you looked to see what's in there? How did you determine that it's grown over 6GB in 3hrs in /opt/zimbra?

[jwilke]

Hi Bill,

Zimbra reports disk usage over 80% over the last few hours it grew from 80 to 94% so thats where the 6gb came from.

Right now things seem to have stabilized though.

redolog archives is filled with 100mb files dated today between 10:10 and 11:30 Typically it created one every few minutes, stopping at 11:30.

[jwilke]

For now I just ran an incremental backup to clear out the redo log archives

will move opt/zimbra to a larger, and more appropriate location next maintenance.