Fuzzball
Cadet
- Joined
- Mar 18, 2014
- Messages
- 3
The short:
I made some mistakes and I thought I had a good backup of everything important, but I'm missing some stuff in those backups. I have a raidz1 with three 3TB drives. Two drives should be functional, but one is listed as unavailable. I'd like to bring the pool online with the two below /dev/gpt/tank-* drives in at least a read only way to grab the data I really need. Is there a way to force mount it?
The long story:
I have been running FreeBSD 10 and rolling my own NAS on a HP Microserver N40L three Seagate 3TB drives in a raidz1 and a Seagate 3TB drive formatted in UFS. I'm at the physical limit of what I can jam in there and wanted to use FreeNAS for the easy web GUI and updates so I got myself an Ebay special Dell C2100 with the backplane that has three SFF-8484 connectors. One SFF-8484 has a breakout cable connects to the four onboard SATA ports and the other two I bought an SFF-8484 to SFF-8087 cable to connect to an IBM M1015 flashed to IT Mode. I think my problems lie with the backplane and/or the cables as I've used the M1015. I also got my hands on several used HGST 3TB drives that I intended to setup as a raidz2 and migrate my data over, but have not been used in this so I won't be referring to them again. Mistake 1 was not testing the system thoroughly. Mistake 2 as not immediately moving the drives back to the old system when I started to see issues.
I setup FreeNAS 11.1 on the C2100. In the C2100 two of the drives in the pool went on the motherboard connections, the third drive in the pool went on the M1015, the UFS drive went on the motherboard (unmounted), and finally a spare Seagate 3TB went on the motherboard. The tank pool mounted fine with everything online. I did not upgrade the version of the zpool. I made a change to one of the unimportant files in the pool and later saw the pool was in a degraded state with the drive on the M1015 being the culprit. I went ahead and told FreeNAS to replace that drive with the spare on the motherboard (Mistake 3). A day passed and only a few KB was resilvered with no active progress. I did a few reboots, but nothing was moving along. I then moved the all the drives over to the M1015 spots with no change (Mistake 4). Then I saw one of the two good drives marked as unavailable and my heart sank knowing I had screwed up.
Here's where things stand:
Both the FreeBSD 10 and FreeNAS 11.1 installs are on separate USB drives so I can move them around easily. I really appreciate it.
I made some mistakes and I thought I had a good backup of everything important, but I'm missing some stuff in those backups. I have a raidz1 with three 3TB drives. Two drives should be functional, but one is listed as unavailable. I'd like to bring the pool online with the two below /dev/gpt/tank-* drives in at least a read only way to grab the data I really need. Is there a way to force mount it?
Code:
$ zpool status tank pool: tank state: FAULTED 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 Dec 23 17:23:11 2017 0 scanned at 0/s, 0 issued at 0/s, 7.30T total 0 resilvered, 0.00% done, no estimated completion time config: NAME STATE READ WRITE CKSUM tank FAULTED 3 0 0 raidz1-0 DEGRADED 43 33 0 7893811221329945986 UNAVAIL 0 0 0 was /dev/gpt/tank-0-K790 gpt/tank-1-K7EW ONLINE 0 0 0 5490429709276012074 UNAVAIL 0 0 0 was /dev/gptid/d8d8a354-e82f-11e7-ace6-60eb699bdba8
The long story:
I have been running FreeBSD 10 and rolling my own NAS on a HP Microserver N40L three Seagate 3TB drives in a raidz1 and a Seagate 3TB drive formatted in UFS. I'm at the physical limit of what I can jam in there and wanted to use FreeNAS for the easy web GUI and updates so I got myself an Ebay special Dell C2100 with the backplane that has three SFF-8484 connectors. One SFF-8484 has a breakout cable connects to the four onboard SATA ports and the other two I bought an SFF-8484 to SFF-8087 cable to connect to an IBM M1015 flashed to IT Mode. I think my problems lie with the backplane and/or the cables as I've used the M1015. I also got my hands on several used HGST 3TB drives that I intended to setup as a raidz2 and migrate my data over, but have not been used in this so I won't be referring to them again. Mistake 1 was not testing the system thoroughly. Mistake 2 as not immediately moving the drives back to the old system when I started to see issues.
I setup FreeNAS 11.1 on the C2100. In the C2100 two of the drives in the pool went on the motherboard connections, the third drive in the pool went on the M1015, the UFS drive went on the motherboard (unmounted), and finally a spare Seagate 3TB went on the motherboard. The tank pool mounted fine with everything online. I did not upgrade the version of the zpool. I made a change to one of the unimportant files in the pool and later saw the pool was in a degraded state with the drive on the M1015 being the culprit. I went ahead and told FreeNAS to replace that drive with the spare on the motherboard (Mistake 3). A day passed and only a few KB was resilvered with no active progress. I did a few reboots, but nothing was moving along. I then moved the all the drives over to the M1015 spots with no change (Mistake 4). Then I saw one of the two good drives marked as unavailable and my heart sank knowing I had screwed up.
Here's where things stand:
Code:
$ ls /dev/gpt freebsd-boot homefs rootfs swapfs tank-0-K790 tank-1-K7EW tmpfs usrfs varfs $ zdb freenas-boot: version: 5000 name: 'freenas-boot' state: 0 txg: 79690 pool_guid: 12110775019371991801 hostname: '' com.delphix:has_per_vdev_zaps vdev_children: 1 vdev_tree: type: 'root' id: 0 guid: 12110775019371991801 create_txg: 4 children[0]: type: 'disk' id: 0 guid: 11021728861805289996 path: '/dev/da0p2' whole_disk: 1 metaslab_array: 33 metaslab_shift: 28 ashift: 9 asize: 32010141696 is_log: 0 DTL: 110 create_txg: 4 com.delphix:vdev_zap_leaf: 31 com.delphix:vdev_zap_top: 32 features_for_read: tank: version: 5000 name: 'tank' state: 0 txg: 30131799 pool_guid: 7103296614445671964 hostid: 3691257932 hostname: 'freenas' com.delphix:has_per_vdev_zaps vdev_children: 1 vdev_tree: type: 'root' id: 0 guid: 7103296614445671964 children[0]: type: 'raidz' id: 0 guid: 17979348125048439281 nparity: 1 metaslab_array: 30 metaslab_shift: 36 ashift: 12 asize: 8995321675776 is_log: 0 create_txg: 4 com.delphix:vdev_zap_top: 160 children[0]: type: 'disk' id: 0 guid: 7893811221329945986 path: '/dev/gpt/tank-0-K790' phys_path: '/dev/gpt/tank-0-K790' whole_disk: 1 DTL: 154 create_txg: 4 com.delphix:vdev_zap_leaf: 161 children[1]: type: 'disk' id: 1 guid: 5811819241200281728 path: '/dev/gpt/tank-1-K7EW' phys_path: '/dev/gpt/tank-1-K7EW' whole_disk: 1 DTL: 156 create_txg: 4 com.delphix:vdev_zap_leaf: 163 children[2]: type: 'disk' id: 2 guid: 5490429709276012074 path: '/dev/gptid/d8d8a354-e82f-11e7-ace6-60eb699bdba8' whole_disk: 1 DTL: 229 create_txg: 4 com.delphix:vdev_zap_leaf: 227 resilver_txg: 30131394 features_for_read: $ zpool status tank pool: tank state: FAULTED 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 Dec 23 17:23:11 2017 0 scanned at 0/s, 0 issued at 0/s, 7.30T total 0 resilvered, 0.00% done, no estimated completion time config: NAME STATE READ WRITE CKSUM tank FAULTED 3 0 0 raidz1-0 DEGRADED 43 33 0 7893811221329945986 UNAVAIL 0 0 0 was /dev/gpt/tank-0-K790 gpt/tank-1-K7EW ONLINE 0 0 0 5490429709276012074 UNAVAIL 0 0 0 was /dev/gptid/d8d8a354-e82f-11e7-ace6-60eb699bdba8 $
Both the FreeBSD 10 and FreeNAS 11.1 installs are on separate USB drives so I can move them around easily. I really appreciate it.
Last edited: