ZFS disk in two pools after reboot?

Question

I'm pretty new to ZFS, so perhaps I'm reading this wrong.
After rebooting my server I had to reimport my pool, and upon doing some was presented with this state:

# zpool status -v
  pool: data
 state: DEGRADED
status: One or more devices could not be used because the label is missing or
        invalid.  Sufficient replicas exist for the pool to continue
        functioning in a degraded state.
action: Replace the device using 'zpool replace'.
   see: http://zfsonlinux.org/msg/ZFS-8000-4J
  scan: scrub in progress since Sat Mar 11 09:54:23 2017
    18.4M scanned out of 28.9T at 4.61M/s, (scan is slow, no estimated time)
    0 repaired, 0.00% done
config:

        NAME                     STATE     READ WRITE CKSUM
        data                     DEGRADED     0     0     0
          raidz1-0               DEGRADED     0     0     0
            sdd                  ONLINE       0     0     0
            5824186382562856058  FAULTED      0     0     0  was /dev/sdb1
            sde                  ONLINE       0     0     0
          raidz1-1               ONLINE       0     0     0
            sdj                  ONLINE       0     0     0
            sdk                  ONLINE       0     0     0
            sdl                  ONLINE       0     0     0
          raidz1-2               ONLINE       0     0     0
            sdg                  ONLINE       0     0     0
            sdb                  ONLINE       0     0     0
            sdf                  ONLINE       0     0     0
          raidz1-3               ONLINE       0     0     0
            sdc                  ONLINE       0     0     0
            sdh                  ONLINE       0     0     0
            sdi                  ONLINE       0     0     0

What grabbed my attention was the FAULTED volume in raidz1-0, but what I didn't notice is that it was /dev/sdb, but /dev/sdb is currently in use in raidz1-2!

So I duly exported the pool, forced a labelclear on /dev/sdb and was presented with this status:

# zpool status
  pool: data
 state: ONLINE
status: One or more devices is currently being resilvered.  The pool will
        continue to function, possibly in a degraded state.
action: Wait for the resilver to complete.
  scan: resilver in progress since Sat Mar 11 10:01:16 2017
    6.16G scanned out of 28.9T at 263M/s, 32h3m to go
    2.02G resilvered, 0.02% done
config:

        NAME                       STATE     READ WRITE CKSUM
        data                       ONLINE       0     0     0
          raidz1-0                 ONLINE       0     0     0
            sdd                    ONLINE       0     0     0
            replacing-1            UNAVAIL      0     0     0
              5824186382562856058  UNAVAIL      0     0     0  was /dev/sdb1/old
              sdb1                 ONLINE       0     0     0  (resilvering)
            sde                    ONLINE       0     0     0
          raidz1-1                 ONLINE       0     0     0
            sdj                    ONLINE       0     0     0
            sdk                    ONLINE       0     0     0
            sdl                    ONLINE       0     0     0
          raidz1-2                 ONLINE       0     0     0
            sdg                    ONLINE       0     0     0
            16211591403717513484   UNAVAIL      0     0     0  was /dev/sdb1
            sdf                    ONLINE       0     0     0
          raidz1-3                 ONLINE       0     0     0
            sdc                    ONLINE       0     0     0
            sdh                    ONLINE       0     0     0
            sdi                    ONLINE       0     0     0

I have two questions

1. How did this happen
2. Presumably this means I have a missing disk somewhere? What's the best way to identify it?

---

Additional info: There should be 12 data disks in this server + boot, but blkid is only 11.

# blkid
/dev/sda1: UUID="43AB-B900" TYPE="vfat" PARTUUID="70dbeb11-8d0f-4a90-892b-71ddbfa40614"
/dev/sda2: UUID="31b78e1e-47d2-4835-84f3-52526382626e" TYPE="ext2" PARTUUID="d4385b72-1d3b-4f10-b7be-a47240d0a875"
/dev/sda3: UUID="BW2exB-GVBK-2kYB-O6I3-0Xff-tZsT-1wR3eT" TYPE="LVM2_member" PARTUUID="8298e710-3c27-45f8-bde2-0ca014f61560"
/dev/sdc1: LABEL="data" UUID="1497224562158568852" UUID_SUB="8549439230979948204" TYPE="zfs_member" PARTLABEL="zfs-d160a62f672223cd" PARTUUID="9a5815bb-0c8c-4147-81f7-3c2ed819c856"
/dev/sdd1: LABEL="data" UUID="1497224562158568852" UUID_SUB="8670871889276024405" TYPE="zfs_member" PARTLABEL="zfs-056f7c2c0a7e1d0a" PARTUUID="672f59c7-b6b3-604b-8afd-594bd3b9b5f8"
/dev/sde1: LABEL="data" UUID="1497224562158568852" UUID_SUB="6213246766257863816" TYPE="zfs_member" PARTLABEL="zfs-65908045daba9599" PARTUUID="04785f97-1125-7642-b5c1-9c1a16cda925"
/dev/sdf1: LABEL="data" UUID="1497224562158568852" UUID_SUB="8276492610986556289" TYPE="zfs_member" PARTLABEL="zfs-f8318dd36075cff4" PARTUUID="5d7feebf-8a5f-654b-b2d1-c15691800f44"
/dev/sdh1: LABEL="data" UUID="1497224562158568852" UUID_SUB="1281571628149249275" TYPE="zfs_member" PARTLABEL="zfs-59cc747b1125d66a" PARTUUID="61c60d91-9a85-3b4d-9b99-8df071434a50"
/dev/sdg1: LABEL="data" UUID="1497224562158568852" UUID_SUB="10881622467137806147" TYPE="zfs_member" PARTLABEL="zfs-1a80f12f1e668bbe" PARTUUID="208107b9-ad5f-184c-9178-5db0ebf19a14"
/dev/sdi1: LABEL="data" UUID="1497224562158568852" UUID_SUB="17007441084616526595" TYPE="zfs_member" PARTLABEL="zfs-0a8e6dabd469faca" PARTUUID="e8ed04a8-cde2-6244-902e-6353664af06a"
/dev/sdj1: LABEL="data" UUID="1497224562158568852" UUID_SUB="8620535390437895467" TYPE="zfs_member" PARTLABEL="zfs-97d91e998134d363" PARTUUID="a689a2ff-3b07-ef41-8b9c-cf6361a0e1d1"
/dev/sdk1: LABEL="data" UUID="1497224562158568852" UUID_SUB="17779182602415489900" TYPE="zfs_member" PARTLABEL="zfs-52c3d94733668a22" PARTUUID="42a8072a-e94c-a64d-aa07-dee30f675655"
/dev/sdl1: LABEL="data" UUID="1497224562158568852" UUID_SUB="7227713853040895948" TYPE="zfs_member" PARTLABEL="zfs-cc1406096601d13c" PARTUUID="5481683e-1d8b-4342-9629-3c49f6397075"
/dev/mapper/server--vg-root: UUID="1e3fee5d-d4c8-4971-ae32-23722bbd0688" TYPE="ext4"
/dev/mapper/server--vg-swap_1: UUID="6447b120-e79d-4c9f-8cc6-8eef5e275dfc" TYPE="swap"
/dev/sdb1: LABEL="data" UUID="1497224562158568852" UUID_SUB="16704776748125199400" TYPE="zfs_member" PARTLABEL="zfs-368140a1f4980990" PARTUUID="c131befd-a122-aa45-b710-399233eb08a6"
/dev/sdb9: PARTUUID="4aaed8f3-443e-2e44-8737-94a4d09496aa"
/dev/sdc9: PARTUUID="5f2cb2dd-dddd-154f-a771-8db4f5475fec"
/dev/sdd9: PARTUUID="22968880-24bb-d94a-a50f-13adaaa380bc"
/dev/sde9: PARTUUID="b867fa3f-bda4-cf40-b44c-c76bad4047be"
/dev/sdf9: PARTUUID="b4f79585-6676-de40-81ea-44cf74937b28"
/dev/sdh9: PARTUUID="77f5225f-e0e5-4d4f-8361-d13984807960"
/dev/sdg9: PARTUUID="be9746bc-1eb5-9342-b753-3471ae936d42"
/dev/sdi9: PARTUUID="08675893-d6d3-0b49-bf69-105383040006"
/dev/sdj9: PARTUUID="107df9dc-7ea8-694a-8deb-7a6025b74b86"
/dev/sdk9: PARTUUID="2b2ef8de-da71-a740-aad0-bd2dc1d1c8a7"
/dev/sdl9: PARTUUID="f52efda2-f758-2f47-80ff-318be5db3fca"

Answer 1

Well, you lost a disk which was member of raidz1-0. And after the reboot the /dev/sd[a-m] devices got renamed, as Michael Hampton already mentioned.

ZFS is smart enough to not rely on the /dev/sdx names and can put the pools together based on the ondisk metadata. At this point your raidz1-0was degraded because of the failed drive, that used to be /dev/sdb before you rebooted the server. After the reboot, the failed disk lead to renaming the disks and the disk which belongs to raidz1-2 became /dev/sdb. As ZFS is smart enough, it didn't care and just put the pools together correctly.
At this point, you just should have replaced the failed disk and resilver the raidz1-0 pool.
Instead you degraded your second pool raidz1-2 by removing the healthy disk /dev/sdb, which truly belongs to raidz1-2 and added it to raidz1-0 resulting in resilvering.

You should replace the failed disk and start the resilvering on raidz1-2. The disks will then be very likely renamed again after a reboot.

To identify the failed disk, make traffic on all disks or volumes and see which disk LED is not blinking when you are in front of your server. Don't forget to make traffic on your root partition disk.
Some hardware vendors do have tools or provide a more elegant way to identify the slot in which the failed disk is in.