SOLVED TN 13.0 Degraded disk, replacement shows-no response on "replace"

[Kris Heslop]

Running TN 13.0, had one disk in 9z2 d (8-10TB WD White Labels shucked from MyBooks and 1-12TB WD White Label. One 10 TB disk failwd a Smart Test and went in to replace the drive with another 12TB already in the box.

Checked smart test results in GUI and it shows failed. Error status notes show servo related.

When I go in to replace the drive, the replacement drive shows. I select it and then when I select "Replace" nothing. No errors, no actions posting on the status window, it just stays like I did nothing.

I read some other thread that suggested gpart destroy -F daXX

This returns: arg0 'da10invalid argument.

Screen shots after "selection" or running command shown.

Seems like a bug, but want to check before I complete another bug report.

 

[Samuel Tai]

This is already called out in the Release Notes for 13. Please use the CLI workaround.

 

[Chris Moore]

This is already called out in the Release Notes for 13. Please use the CLI workaround.

How does this "CLI workaround" handle a system with multiple pools?

 

[Samuel Tai]

How does this "CLI workaround" handle a system with multiple pools?

python3 replace_disk.py <pool_name> <gptid/####> <ada#>

As this is just a wrapper script for zpool replace, the affected pool name is an input.

 

[Kris Heslop]

Thanks, that was faster than I expected.

I will look at that further later tonight.

Some questions so far, since no data integrity issues are reporting, I don't have anything matching the gptid with the da#, is it safe to figure it is disk # 7 in the list and count down, e.g. 7th in the pool listing, so 7th gpid down in the shell's listing ? Or can you refresh me where to match the gpid and ada/da#? (Note I have one cache disk that is after the main disks listed in the pool status listing)

I did see the "python3" and from what I read, I have to use the Curl -s to download it for use, and the one question I just wanted to confirm was it was safe to use the da# for the replacement disk, which you reconfirmed .

Thank you Samuel and Chris.

 

[Samuel Tai]

Run zpool status -v <name of your pool> to get the GPTIDs in the pool. It will tell you which one is not ONLINE. To verify that GPTID is the correct disk, run glabel status, and it will match the GPTID to a port ID (e.g., daX).

Also look at Storage->Disks for the serial number of the affected drive. Make sure you pull the correct disk; match it against the serial number.

 

[Chris Moore]

Or can you refresh me where to match the gpid and ada/da#?

You might also want to get the script described here:

SOLVED - Utility: disklist.pl, for listing partition, gptid, slot, devices, disktype, serial num, & multipath

A contributor on the Serve the Home forum contributed a wonderful utility. Sébastien (Username nephri) has written a wonderful perl script https://forums.servethehome.com/index.php?threads/freenas-script-for-a-convenient-way-to-report-installed-disks.16310/ reports installed disks I am...

www.truenas.com

Also a pearl script, it give drive information for all drives in the system. I find it very useful.

 

[Kris Heslop]

Chris, I really liked the idea of that script, but I am not sure that I'm that literate... So I used the one Samuel shared...and it gave me a bunch of lines...that ended with the gptid being wrong...and once I got the "b"s that I wrote down as 6's fixed I got nothing...and I thought if it is resilvering, it will take a while to come back...

So I checked on the pool status page and it shows the (two) disks as "replacing" so success.

Samuel and Chris, thanks for directing me to the right place.

BTW, While I do read the release notes, that one would not have likely connected for me, even if I reread it, since it says it "fails with javascript error." and I got nothing at all.