Last night, I found that I had a number of unused ZFS snapshots. I decided to destroy some of them. The first destroy locked up the system. Nagios reported everything was dead. I remotely power cycled the system. After it came back, I noticed that a scrub was in progress. This had not been initiated because of the power cycle. The scrub was a weekly automatic operation which started some hours before my destroy attempt.
I decided to let the scrub complete before proceeding.
Here is what I stared with:
$ zpool status
pool: storage
state: ONLINE
scan: scrub repaired 0 in 20h24m with 0 errors on Wed Oct 3 23:38:44 2012
config:
NAME STATE READ WRITE CKSUM
storage ONLINE 0 0 0
raidz2-0 ONLINE 0 0 0
gpt/disk01-live ONLINE 0 0 0
gpt/disk02-live ONLINE 0 0 0
gpt/disk03-live ONLINE 0 0 0
gpt/disk04-live ONLINE 0 0 0
gpt/disk05-live ONLINE 0 0 0
gpt/disk06-live ONLINE 0 0 0
gpt/disk07-live ONLINE 0 0 0
errors: No known data errors
$ zfs list
NAME USED AVAIL REFER MOUNTPOINT
storage 7.99T 911G 14.5G /storage
storage/compressed 7.96T 911G 53.3K /storage/compressed
storage/compressed/bacula 7.59T 911G 4.66T /storage/compressed/bacula
storage/compressed/bacula@2011.03.21 2.93T - 3.46T -
storage/compressed/home 2.75G 911G 2.75G /storage/compressed/home
storage/compressed/jails 373G 911G 373G /storage/compressed/jails
storage/pgsql 13.8G 911G 13.8G /storage/pgsql
$ zfs list -t snapshot
NAME USED AVAIL REFER MOUNTPOINT
storage@2011.03.21 93.8K - 14.5G -
storage/compressed@2011.03.21 38.4K - 51.2K -
storage/compressed/bacula@2010.10.19 8.90G - 2.29T -
storage/compressed/bacula@2010.10.20 159K - 2.29T -
storage/compressed/bacula@2010.10.20a 78.9K - 2.29T -
storage/compressed/bacula@2010.10.20b 78.9K - 2.29T -
storage/compressed/bacula@pre.pool.merge 1.12M - 2.58T -
storage/compressed/bacula@2011.03.21 360G - 3.46T -
storage/compressed/home@2011.03.21 122K - 2.75G -
As you can see, I have some snapshots which are nearly two years old. I don’t need them any more. Let’s let them go.
# zfs destroy storage/compressed/bacula@2010.10.20 # zfs destroy storage/compressed/bacula@2010.10.20a # zfs destroy storage/compressed/bacula@2010.10.20b # zfs destroy storage/compressed/bacula@pre.pool.merge # zfs destroy storage/compressed/bacula@2010.10.19 # zfs destroy storage@2011.03.21 # zfs destroy storage/compressed@2011.03.21 # zfs destroy storage/compressed/home@2011.03.21 # zfs destroy storage/compressed/bacula@2011.03.21
Now I have:
$ zfs list NAME USED AVAIL REFER MOUNTPOINT storage 5.06T 3.82T 14.5G /storage storage/compressed 5.03T 3.82T 53.3K /storage/compressed storage/compressed/bacula 4.66T 3.82T 4.66T /storage/compressed/bacula storage/compressed/home 2.75G 3.82T 2.75G /storage/compressed/home storage/compressed/jails 373G 3.82T 373G /storage/compressed/jails storage/pgsql 13.8G 3.82T 13.8G /storage/pgsql
I’ve gone from 8T to 5T in use. Which means I have 3T to grow. That’s enough for now. :)
Here is a graph from munin which shows the change in disk consumption resulting from my Bacula and ZFS efforts over the past two days.
