Recommendation on degraded 8.0.4 system

[kesh]

Hi,

I set up FreeNAS 8.0.4.a long time ago when the version came out and left it as is since. Last night, I got an "degraded system" alert notifying one of the ZFS volume is dead.

Should I upgrade it to the latest version before replacing the failed drive?

Thanks,
Kesh

 

[HoneyBadger]

You should get your zpool back to a healthy state first before doing any upgrades; or even if you aren't planning on upgrading.

 

[kesh]

Thanks for the tip, HB!

Interestingly though, the alert went away when I rebooted; the reported drive came back to life... I'm going to upgrade to the latest version now and keep my eyes on it.

 

[HoneyBadger]

Thanks for the tip, HB!

Interestingly though, the alert went away when I rebooted; the reported drive came back to life... I'm going to upgrade to the latest version now and keep my eyes on it.

I'd run a SMART test on the drive - or at least pull the attributes - and look for any signs of a failure yet to come. You might be lucky in that it's giving you a warning as opposed to throwing in the towel without notice.

Good luck with the upgrade!

 

[gpsguy]

Don't upgrade straight to the latest version. Consider going from 8.0x to 8.2x to 8.32 them onto 9.x. Backup your config along the way and don't upgrade your pool until your are sure of the changes. You can download older versions at: http://download.freenas.org


Sent from my phone

 

[cyberjock]

Thanks for the tip, HB!

Interestingly though, the alert went away when I rebooted; the reported drive came back to life... I'm going to upgrade to the latest version now and keep my eyes on it.

That wasn't exactly the best choice. You should have investigated what went wrong. Any evidence that would have helped was lost on reboot. But too late to cry over spilt milk.

Definitely do some scrubs and SMART tests though.

 

[joeschmuck]

Since we have no idea what your system configuration is nor how you use FreeNAS you might be better off sticking with 8.0.4. As everyone has suggested, run some SMART tests, especially the Long tests on all the drives. Could you post the results of 'smartcrl -a /dev/adax' (within code brackets) where adax equals the drive, for each drive of course. Looking at ID values 5 and 197 will help indicate if there are also possible concerns.

As for upgrading, there actually are specific steps you must make to ensure your database remains in-tact. Since you have 8.0.4 then you likely do not have a complicated configuration and could write it down, then just load up the 9.2.1.3 image and make changes to your settings. I certainly would not go through all the updates to get to the end, it would take too long, just a waste of time. If you do not have 8GB of ECC RAM then you should not upgrade.

Good Luck.