Replacing failed drive

[markgca]

im new to this, so...
have a freenas 11.3-U2 running fine, but a infant mortality drive dropped out after many errors
got a replacement, no room to put in a spare drive

powered the box down, removed the damaged drive, replaced it with a new one, powered up

told it to replace the drive, and it is resilvering now. but...it doesnt look right, so perhaps i did something wrong (like maybe i was supposed to detach the drive before i powered down? not sure)

anyway, here is a screenshot; any guidance appreciated.

note that it appears to have two drives under replacing, one is offline (the one i removed?) and the new one is where it should be.
my concern is that it still shows the one i removed, or is that not what im seeing/not an issue? i rebooted to make sure it wasnt that

 

[psssttt]

Hi @markgca ,

It looks like you're facing the same problem as I do.
Here's my post with things I've done so far:

SOLVED - After replacing a disk and resilvering pool is still in degraded status and drive says it's still replacing

Hi, In recent weeks two (/dev/ada3 & /dev/ada5) of my eight drives started to report more and more bad sectors and errors so I've decided to replace them. I followed official instructions for replacing a failed disk...

www.ixsystems.com

Cheers

 

[sretalla]

It looks like you're facing the same problem as I do.

I don't think the issue is the same... the screenshot clearly shows that the OP is in the process of resilvering, which is just not done yet... the replacing disk pair will disappear when the process is done and you can check with zpool status at CLI if you wish. This is a flaw in the process that the status screen in the GUI doesn't accurately reflect the resilver progress.

Here's my post with things I've done so far:

I'm a little less than keen to dive into your post as you're digging yourself into a much deeper hole than necessary after experiencing some file corruption (which seems to me to have been the only real problem), which you could have corrected by dealing with those corrupted files either by restore from backup or simple deletion. The resilver will never finish properly with corrupt files in the pool.

You may want to look into the SMART status of all of your disks too, since corruption doesn't just come from nowhere.

 

[psssttt]

Thank you @sretalla for the info on scrub & corrupted files.
I don't want to start convo re. my issue here, but does it mean that when a scrub finds corrupted files then replacing a failing drive is pointless?

 

[sretalla]

For a resilver to complete, it needs to be able to confirm that the data on the replaced disk is consistent... with corrupted files in the pool, that will never happen.

A removed/replaced disk should not imply that files are/were corrupted and is not normally the case.

Just deal with the corrupted files and it will be fine.

 

[PhiloEpisteme]

11.3-U2

Once you're all done with the resilver and get your pool back to a steady state you should upgrade to 11.3-U2.1 immediately. There is a nasty bug which can cause data loss in earlier versions of 11.3. See the release notes for more detail.

 

[markgca]

ok, so it took forever but eventually cleared and now the vdev is showing online. now the only issue i have is that the replacement drive self healed number did not reset; is there something i need to do here?

edit: never mind, i updated to -U2.1 as recommended and it cleared (likely reboot? )
My first disk failure in freenas and i was a bit nervous, but it turned out great; i guess those 16tb drives just take forever to resilver

 

[psssttt]

How did it go @markgca ?

 

[markgca]

How did it go @markgca ?

took forever, but i guess a 16tb drive didnt help. and everything is good
however; during the resilver it was scary, first with that original issue i posted, second if you clicked on dashboard/pool/tank NONE of the drives showed up, and rather had something on the order of "chevron " on each line; i assume that might be an icon of some sort that didnt show up for the entire time it was resilvering. that came back after completion, just another stressor that something was going wrong. likely a bug that im not going to try to reproduce

 

[Yorick]

Out of curiosity: What's "forever"? How long did it take?

 

[markgca]

couple days; up until freenas i used mirrored arrays which replicated quickly.