Thanks for using the TrueNAS Community Edition issue tracker! TrueNAS Enterprise users receive direct support for their reports from our support portal.

Spare not being added to pool upon degredation

Description

TrueNAS-12.0-INTERNAL-199

Pulled drive from mirrored pool (with spare). Pool degrades, but spare never gets activated/used as a part of the pool. See screenshots.

Problem/Justification

None

Impact

None

Activity

Show:

Ryan MoellerJuly 20, 2021 at 6:39 PM

Jeff reported that everything went smoothly when he tried again using a set of matched disks for the pool and spare.

It would be nice to see middleware do the partitioning the same for both sector sizes so that a 4k native disk can replace a 4k/512b disk of the same size, but that would mean partitioning the 4k/512b disks differently, then they wouldn't be able to replace an identical disk, which is worse.

Ryan MoellerJuly 20, 2021 at 6:01 PM

Assuming the removed drive is identical to its mirror:

 

root@tn02a[~]# glabel status Name Status Components gptid/e48415f2-e8a6-11eb-9222-3cfdfe4f81c8 N/A nvd0p1 gptid/8e9f2c5e-e8ba-11eb-81b5-3cfdfe4f81c8 N/A da23p1 gptid/0fbe3ab0-e972-11eb-81b5-3cfdfe4f81c8 N/A da3p1 gptid/880b67dd-e972-11eb-81b5-3cfdfe4f81c8 N/A da4p1 gptid/88132d5e-e972-11eb-81b5-3cfdfe4f81c8 N/A da5p1 gptid/ad8b3095-e972-11eb-81b5-3cfdfe4f81c8 N/A da6p1 root@tn02a[~]# gpart show da6 => 6 1953506635 da6 GPT (7.3T) 6 122 - free - (488K) 128 1953506513 1 freebsd-zfs (7.3T) root@tn02a[~]# gpart show da3 => 40 15628053088 da3 GPT (7.3T) 40 88 - free - (44K) 128 15628053000 1 freebsd-zfs (7.3T)

da3 is the remaining drive in mirror-0, da6 is the spare. They have been partitioned differently, even though they are the same size in bytes:

root@tn02a[~]# diskinfo -v da3 da3 512         # sectorsize 8001563222016 # mediasize in bytes (7.3T) 15628053168 # mediasize in sectors 4096         # stripesize 0           # stripeoffset 972801       # Cylinders according to firmware. 255         # Heads according to firmware. 63           # Sectors according to firmware. HGST HUH728080AL5200 # Disk descr. 2EKJL84X     # Disk ident. id1,enc@n5b0bd6d1a303ecbd/type@0/slot@4/elmdesc@slot03_______ # Physical path No           # TRIM/UNMAP support 7200         # Rotation rate in RPM Not_Zoned   # Zone Mode root@tn02a[~]# diskinfo -v da6da6 4096         # sectorsize 8001563222016 # mediasize in bytes (7.3T) 1953506646   # mediasize in sectors 0           # stripesize 0           # stripeoffset 121600       # Cylinders according to firmware. 255         # Heads according to firmware. 63           # Sectors according to firmware. HGST HUH721008AL4205 # Disk descr. 7SHP1G9W     # Disk ident. id1,enc@n5b0bd6d1a303ecbd/type@0/slot@7/elmdesc@slot06_______ # Physical path No           # TRIM/UNMAP support 7200         # Rotation rate in RPM Not_Zoned   # Zone Mode

Different sector size is the likely culprit, but ideally we should be partitioning in a way that is tolerant to this.

 

 

Ryan MoellerJuly 20, 2021 at 4:42 PM
Edited

zfsd says the spare is too small to replace the device:

root@tn02a[~]# grep zfsd /var/log/messages Jul 20 08:57:08 tn02a 1 2021-07-20T15:57:08.036051+00:00 tn02a.qe.ixsystems.net zfsd 1191 - - Replace vdev(test1/16676033406322047726): cannot replace 16676033406322047726 with /dev/gptid/ad8b3095-e972-11eb-81b5-3cfdfe4f81c8: device is too small
Behaves as Intended

Details

Assignee

Reporter

Components

Fix versions

Affects versions

Priority

More fields

Katalon Platform

Created July 20, 2021 at 4:07 PM
Updated July 6, 2022 at 8:57 PM
Resolved July 20, 2021 at 6:17 PM

Flag notifications