Solaris 11.3 brings ZFS task monitoring enhancement – zpool monitor

With the Solaris 11.3 beta release Oracle added a beautiful zpool command. Zpool monitor! And who would have figured, it does exactly that. It helps you to keep track of send, receive, scrub and/or resilver actions.

Let’s say you run zpool scrub rpool and you are interested in the status/progress. So far zpool status was the command of choice to do so. Usually used somehow like while :; do zpool status …|grep …; done. Works but not really convenient. And just imagine have a few send and/or receives running.
Well it is all taking care of. zfs monitor will the job for you. A long needed zfs feature. Thanks a lot Solaris engineering. Very nice improvement.

So how does zpool monitor work. Let’s take a look:

root@wacken:~# zpool help monitor
        monitor -t provider [-T d|u] [[-p] -o field[,...] [pool] ... [interval [count]]
        Valid values for 'provider' are send, receive, scrub, and resilver
        Valid values for 'field' are done, other, pctdone, pool, provider, speed, starttime,
          tag, timeleft, timestmp, total

This is what the default looks like for monitoring scrub:

root@wacken:~# zpool monitor -t scrub 1
pool                            provider  pctdone  total speed timeleft
rpool                           scrub       0.0    64.8G 1.66M 11h08m
rpool                           scrub       0.0    64.8G 1.69M 10h55m
rpool                           scrub       0.0    64.8G 1.71M 10h45m
rpool                           scrub       0.0    64.8G 1.74M 10h36m
rpool                           scrub       0.0    64.8G 1.79M 10h19m
rpool                           scrub       0.0    64.8G 1.80M 10h15m
rpool                           scrub       0.8    64.8G 27.3M 40m09s
rpool                           scrub       0.8    64.8G 26.5M 41m23s
rpool                           scrub       1.0    64.8G 31.9M 34m23s
rpool                           scrub       1.1    64.8G 33.6M 32m34s
rpool                           scrub       1.1    64.8G 32.2M 33m58s
rpool                           scrub       1.1    64.8G 30.9M 35m22s
rpool                           scrub       1.1    64.8G 29.8M 36m44s

Let’s add a few more fields to it.

root@wacken:~# zpool monitor -t scrub -o pool,provider,pctdone,total,speed,timeleft,tag,starttime,done 1
pool                            provider  pctdone  total speed timeleft   tag                 starttime done
rpool                           scrub       3.5    64.8G 18.8M 56m51s     scrub-267           14:22:58  0
rpool                           scrub       3.5    64.8G 18.7M 57m05s     scrub-267           14:22:58  0
rpool                           scrub       3.5    64.8G 18.6M 57m29s     scrub-267           14:22:58  0
rpool                           scrub       3.5    64.8G 18.5M 57m52s     scrub-267           14:22:58  0

Ain’t this nice. If you have an environment that includes automated zfs snapshots together with send/receive this will really make your day.
This command will definately get an alias in my zshrc ;-)

Leave a Reply