Bug 224272 - zfs replace doesn't detach old device
Summary: zfs replace doesn't detach old device
Status: New
Alias: None
Product: Base System
Classification: Unclassified
Component: kern (show other bugs)
Version: 11.0-RELEASE
Hardware: Any Any
: --- Affects Some People
Assignee: freebsd-fs (Nobody)
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2017-12-12 12:56 UTC by Matthias Pfaller
Modified: 2017-12-13 07:41 UTC (History)
1 user (show)

See Also:

Attachments
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Matthias Pfaller 2017-12-12 12:56:20 UTC 
One of our drives is starting to show errors with smart self tests. Output of zpool status:

  pool: zmnt
 state: ONLINE
  scan: resilvered 1.03T in 14h11m with 0 errors on Mon Dec 11 22:23:44 2017
config:

        NAME          STATE     READ WRITE CKSUM
        zmnt          ONLINE       0     0     0
          raidz2-0    ONLINE       0     0     0
            gpt/zfs0  ONLINE       0     0     0
            gpt/zfs1  ONLINE       0     0     0
            gpt/zfs2  ONLINE       0     0     0
            gpt/zfs3  ONLINE       0     0     0
            gpt/zfs4  ONLINE       0     0     0
            gpt/zfs5  ONLINE       0     0     0
        logs
          mirror-1    ONLINE       0     0     0
            gpt/zil0  ONLINE       0     0     0
            gpt/zil1  ONLINE       0     0     0
        spares
          gpt/zfs6    AVAIL   
          gpt/zfs7    AVAIL   

errors: No known data errors

I tried to replace gpt/zfs4 with gpt/zfs6:
  zpool replace zmnt /dev/gpt/zfs4 /dev/gpt/zfs6

After the resilver finished the pool was in the same state as before...
Comment 1 eborisch+FreeBSD 2017-12-12 23:34:24 UTC 
Perhaps because it is already marked as a spare (part of the pool) and the ZFS later hasn’t had any errors yet? zpool remove the spare and then re-issue the replace?
Comment 2 Matthias Pfaller 2017-12-13 07:41:05 UTC 
I already did what you suggested and the system did what I wanted to achieve. 
Its still surprising to watch it resilver for 14h and then ending up in the same state as before...
I think when you request a replace (even when replacing with an hot spare) it should just detach the replaced drive when finished.
So maybe that's not a bug report but a feature request.