Pool Offline, cannot import, but shows up Online via zpool import

[kooplaah]

So miraculously... I was able to import the pool!!! This is going to sound probably very derivative to anyone that knows zfs properly, but once I was able to ID the faulted drives I just pulled the 3 out, threw in the 1 spare I had, ran a zpool replace and then ran the import command, and it worked! It resilvered the pool and one reboot later the pool was back!

I did also upgrade the HBA firmware prior to that, so I'm hoping this solves the corruption issue... I was thinking more about it after reading that other forum post and I did run this exact hardware with Unraid for a while with no issue. If there was a specific TruNAS incompatibility it would explain why I never had this issue when running Unraid.

So... small victory but I'm very nervous still. Rookie question, but how can I be sure I have this backed up in case a failure happens again? I happen to have 3 slots open on my HBA now and I have 3 other drives I'm planning to run more permanently as a backup pool. Can I throw these in there, create a new pool for these 3, and do a replication? I'm not super clear on all of the TruNAS options in the Data Protection area... I can probably do some research myself to get a better idea, but any suggestions are welcome.

 

[sfatula]

Well, that would be a local backup but doesn't save you in the event the system gets stolen, fire, etc. Given your issues, it's a good start though, so, yes you can do as you say, but that is not the only backup you should do.

I backup to offsite ZFS using replication. I further use the backup software named Kobia to also back up to a remote site at file level for the most important stuff, and this doesn't require zfs offsite. I further do local backups of various sorts, one of which is replication to external USB drives which I rotate to offsite storage, my bank safe deposit box every few months. So, one local backup set is on site, one is at the bank.

So happy it worked out!

I doubt it was Truenas that had an incompatability per se. Most likely was a driver issue, maybe unraid was differnet kernel, etc. I recall having some sort of driver issue bug in debian related to the LSI controllers.