Strange spool after HD firmware upgrade

[adrianwi]

I'd hijacked a how-to thread after a strange problem with my zpool following a hard drive firmware upgrade and was asked to post as a new thread, so here goes!

Original how-to thread here > http://forums.freenas.org/index.php?threads/updating-seagate-hard-drive-firmware.14561/#post-102194

Firmware upgrade appeared to go OK on two drives, but no firmware available for the other two even though FreeNAS is giving me a warning that one may be available.

I then got a critical error e-mail but was at work so couldn't investigate > The volume APE_pool (ZFS) status is DEGRADED

This is the post I made after getting home to investigate:

Very strange! I'm still pretty new to this, so really not sure exactly what I've done but it seems to have cured the problem.

When I checked the Volume Status the RAIDZ1 was showing the 4 disks with one without the ada1p2 and a long text string and UNAVAIL next to it. I assumed this was a faulty drive, but unsure which one to remove did a little more digging in the WebUI. Couldn't seem to work it out, so went back to Volume Status and clicked the only option available when clicking the offending disk - Replace (and not Remove as originally posted)

Clicked this and it showed me the serial number, so thought I'd be able to shutdown and then identify the disk. After clicking replace (and not remove as originally posted), I got a new option - Detach. As I was about to click this, I noticed that I now had all 4 disk with ada0-3p2 showing all with a status of ONLINE. Thought I'd click Detach on the UNAVAIL one and see what happened - it just disappeared.

I now have a yellow warning light in the WebUI telling me that one or more of the devices is being resilvered so it looks like it's fixed itself.

I've ordered a replacement drive, just in case, so will just keep an eye on things.

Any idea what might have happened?

ps: just noticed auto-correct has changed title to spool instead of zpool!

 

[cyberjock]

No clue. Your comment is very incoherent. What you claim to have pressed and what you pressed don't really meet up with reality.

Normally, you'd have to offline the disk, then detach it, then replace it. You have no comment at all about pressing "replace" so I can't help you at all. The only thing I can think of is that you replaced the disk while randomly clicking stuff. Randomly clicking isn't something anyone should do. It's cause data loss galore for many users. Next time you might want to consult the manual as you might not be so lucky next time.

 

[adrianwi]

Apologies if the description isn't very clear, as I was trying to describe it after it had happened. Whilst on this occasion I'd also accept I was a little more gung-ho than usual (I'd backed up all my data prior to attempting the firmware update, so the only risk was time), I had still read the manual regarding swapping out devices but it didn't really describe what actually happened. Under normal circumstances, I'd have asked for further advice before 'randomly clicking' anything!

After giving it some more thought, I think the zpool took one of the drives offline and detached automatically, which would explain why the device showing as UNAVAIL only had the option to Replace (not Remove as I'd typed above). I'm assuming by clicking replace, it then added this device back into the zpool and started the resilver process.

Still no idea why a firmware upgrade would result in a device being removed and detached automatically, but as everything was easily solved I'm not too bothered at this stage. The resilvering was about 45% complete when I left for work, so will hopefully complete later this evening.

Thanks anyway for your somewhat condescending response :D

 

[cyberjock]

That makes much more sense. If the disk had gone UNAVAIL the only option would have been "REPLACE". Then ZFS would resilver the disk back into the pool. That sound very logical.

As for the firmware update, no clue. Not a fan of Seagate anymore and I don't try to do firmware updates on hard drives unless I know there's a fatal flaw in the firmware.

 

[adrianwi]

Reason for the firmware upgrade was that FreeNAS wouldn't run any SMART tests due to the firmware. It appears the upgrade has resolve this and the resilver is now complete and everything looks back to normal.

 

[cyberjock]

I've never heard of a firmware upgrade having anything to do with SMART tests.. Can you provide a link to this? I'm very intrigued...

 

[adrianwi]

I could have done before updating the firmware, but I'm not getting the error message now telling me that smartctl can't be run.

I'll check my logs to see if there's anything in there...