Can I clone a failing member disk and have it replace the disk in the pool

[Maddog3115]

I have a Z2 array and for some reason I didn't get any smart/degraded messages from PVE until the 3rd disk started getting SMART errors? I replaced the failed drives but I need the latest drive (with read errors) for the pool to function. Can I DD the failing drive to replace it in the pool so that I can resilver to the other 2 replaced drives? Sorry fairly new to Proxmox.

Also all the drives were all the same model so not sure why I didn't get smart/degraded messages sooner.

Thanks,

 

[Dunuin]

Check with smartctl -a /dev/YourDisk what the "UPDATED" value of your smart attributes are. Consumer drives often got "offline" there and will only update the values when you reboot the server. Thats fine for a client machine / workstation thats get shutdown over night, but can be problematic when running a server that runs months or years without rebooting (so you only see bad values when its too late). Enterprise disks usually got "Always" there, so they can update smart values while running.

Also check that your ZFS pool gets scrubbed regularily (for example zpool scrub YourPool once per month). And setting up automated smart tests using cron can be useful too (something like smartctl -t short /dev/yourDisk or smartctl -t long /dev/yourDisk).
And make sure you setup postfix and zfs-zed so you get email notifications in case your pool degrades.

 

[Maddog3115]

Thanks for the reply. I will have to make those changes if I can get this back online reliably. Will I be able to do the clone to get the Pool back online long enough to resilver to newer drives? I've attached a screenshot of the last status of the Pool, and the second online drive is the one reporting SMART errors and read errors in syslog on PVE