SOLVED HELP!!! Accidentally shutdown at the beginning of resilver

[Havock2]

I replaced one of my drives and told the machine to resilver. I accidentally hit the power button while replacing the covers. This started shutting the machine down. Now, it won't boot. It tries to scan my volumes and suspends them. After this is done, it reboots.

I tried putting the original drive back in and that didn't help. I also checked that all cables are secure.

 

[DrKK]

This is really very intriguing.

There really should be no ill effect from a proper ACPI shutdown in the middle of a resilver.

Are you leaving anything out of the story? We're discussing it in the IRC channel right now and we're all at a loss.

 

[DrKK]

Also, we'd like to know what "it suspends my drives" means.

 

[Havock2]

I waited for the machine to fully power off before I turned it back on.

I mis-spoke about suspending my drives. It tries to import the ZFS volume. Then it tries to load each dataset and zvol. This fails with a message about suspending. I can't see the whole line, because my console cuts it off.

This was VERY early in the resilver. When it shutdown, could it have been re-doing the partition tables?

 

[Havock2]

Is there a way to skip the import of the ZFS volume and pull logs?

 

[DrKK]

Will the system start with no drive, whatsoever, in for the "broken" drive? You said it was a RAID-Z3 in your signature, so the pool should be viable.

 

[Havock2]

That did not help.

 

[DrKK]

Well, sir, unfortunately, there appears to be some piece of information here that we're lacking. Certainly ZFS will not destroy your pool based on the power going down in the middle of a resilver, especially a graceful ACPI shutdown, which is what you describe.

I guess the next thing to try is a fresh boot device, with a freshly installed freenas. This will have the benefit of being known-clean, and it will not expect nor try to mount your pool. This can either be temporary boot device for trouble shooting, or permanent, as a replacement. Then we boot into this with no pool expected, which I am assuming (barring a hardware error) it will do, and we see where we are with respect to zfs reporting if it will be import your pool, and/or what camcontrol devlist reports.

 

[Havock2]

I finally have the chance to work on this again. I installed FreeNAS on a spare drive that I had laying around.

It's now trying to import my volume. It has been taking a while to import. Is this normal?

I had the Watch Dog enabled in my BIOS. Could this have been my problem?

 

[Havock2]

The volume successfully imported into my new install. Can I just put my old drives back in place and try to boot?

 

[DrKK]

I am afraid I am bit confused. If you were able to import your pool, then presumably the drives were back in place, so what does "put your old drives back in place" mean?

Also, watchdog won't be the problem. FreeNAS runs watchdogd automatically out of the box to handle that.

No, it is not normal to take a long time to import, but I am wondering if you had some transactions roll back or something. Future imports should not be too bad.

 

[Havock2]

Sorry about the confusion. I was able to successfully import the pool in the new install. It just took a while. I can see all of my datasets and zvols. It's resilvering now. I'm going to let this finish before I do anything else.

After that, I'm wondering if my next step is to put my production boot drives back in place and try to boot.

 

[DrKK]

ah!!! Now I understand.

Should you put the BOOT drives back in.

That should be alright. If the boot devices are screwed up, you'll know soon enough, and then can reinstall the FreeNAS on them easily enough.

Smart move letting the resilver finish.

You are doing The Right Things.