draid replace routine?

[kbuggenhout]

System information

Type	Version/Name
Distribution Name	CentOS
Distribution Version	7.3
Linux Kernel	3.10.0-514.26.2.el7
Architecture	x86 64 bit
ZFS Version	0.7.0-rc3
SPL Version	0.7.0-rc5

Describe the problem you're observing

creating a draid2 vdev on 4 drives works like a charm
but when I do the test offline 1 drive , the spare kicks in as expected.
when I online the drive again it does not free the spare

when I remove the drive from the pool, ( zpool detach tank sdc) and use
zpool replace -f tank '$draid2-0-s0' sdc it does its job, but now I am running without a spare.

guess I should not detach the drive from the pool.

Describe how to reproduce the problem

see above... offline a drive then detach and replace the spare with that free drive

Include any warning/errors/backtraces from the system logs

no warnings or errors, fs still online

[thegreatgazoo]

Let's say sda has failed, and replaced by $draid2-0-s0. Now normally what the admin would do is to replace the failed sda with a new drive, using zpool replace. Once that's done, the $draid2-0-s0 will become available again.

In your case, the offline drive came back online again, if the drive is again healthy, you can simple detach the $draid2-0-s0. That'd make the dspare available again, which essentially undo the rebuild.

[kbuggenhout]

will try again failing a drive by disconnecting it logically, then after sync build bring it back in, do the replace drive procedure, and see how it reacts... thnx..

any idea if this is getting merged into the tree some day?

[thegreatgazoo]

Please try with the latest code zfsonlinux#7078