Monitoring ZFS

[decibel83]

Hi,
when I was using hardware RAID controllers on old servers I was using the RAID controller commands (megacli for LSI, arcconf for Adaptec and so on) to monitor array and drives status.

What I should use for ZFS?

Are smartctl, zfs-zed and zpool status enough to be alerted and to predict hardware and pool failures?

Thanks!

 

[guletz]

Hi,
when I was using hardware RAID controllers on old servers I was using the RAID controller commands (megacli for LSI, arcconf for Adaptec and so on) to monitor array and drives status.

What I should use for ZFS?

Are smartctl, zfs-zed and zpool status enough to be alerted and to predict hardware and pool failures?

Thanks!

Only smartctl can predict a future failure of a disk. In my opinion I see very often that WD can not be realistic when a disk will fail (no useful smartd info in this case). But for Seagate and HGST, I was see many cases when smartd was very useful.
At the zfs layer I also can say (in my opinion ) that in many cases erors after a scrub has indicated a pre-failure event. zfs-zed can email you when zfs has problems.

 

[fireon]

Zpool see problems with the pool. So when a HDD is damaged then you see an error. For Nagios there are some checkplugins. With CheckMK ZFS-Pools were montiored automaticly, also the smarterrors.

 

[LnxBil]

We monitor zpool status -v and last scrub date. Works fine.