ehauk Posted September 12, 2016 Share Posted September 12, 2016 So I had my first parity check error (only 1 error) a few days ago. One of my disks has a Current pending sector = 1. I ran an extended smart test and it completed without error. I have a new drive that is currently preclearing. There appears to be different trains of thought as to how to proceed. I'm not sure if I should rebuild with the new disk or rerun the parity check. Any advice would be appreciated. Hitachi_HDS5C3020ALA632_ML2220F30V1YHE-20160912-0940.txt Link to comment
garycase Posted September 12, 2016 Share Posted September 12, 2016 Assuming the parity error was corrected (Do you run correcting checks?), you should confirm all is well by running another parity check. If your check was non-correcting, you need to run a correcting check so the error will be fixed -- THEN run a second check (I always just run correcting checks, but some folks prefer to only do those when they know there's an error). Link to comment
garycase Posted September 12, 2016 Share Posted September 12, 2016 ... r.e. the pending sector => If you have enough space on another system, I'd copy the entire contents of the drive with the pending sector to that drive; then force a reformat of that drive [assign a different file system; Start the array; Stop the array and assign the file system you actually want (e.g. XFS); and then Start the array again and let it format the drive; and then copy the data back. That should force reallocation of the pending sector. If it does not, I'd simply replace the drive. Link to comment
ehauk Posted September 12, 2016 Author Share Posted September 12, 2016 I do have it set to write corrections to parity disk. Does that influence your suggested course of action? I do not currently have enough storage space to copy all the contents to another location. Should I add my new disk to the array and use it to shuffle around the data? Link to comment
trurl Posted September 12, 2016 Share Posted September 12, 2016 Since your parity is good, and you have a new drive preclearing, I would say that if the new drive passes, just use it to rebuild. If that goes OK then you can deal with trying to get the pending sector cleared on the old drive by preclearing it to see if it can be reused. Link to comment
garycase Posted September 13, 2016 Share Posted September 13, 2016 Agree => You can simply replace the drive with the pending sector with your new drive. I'd do the following: (a) Replace the drive with your new drive and let the system rebuild it. [The safest way to do this is Stop the array; unassign the drive; Start the array so it's shown as "missing"; Stop the array and assign your new drive to that slot; and then Start the array and wait for it to rebuild the drive (many hours). (b) Do a parity check to confirm all is well after the rebuild. This is the ONE time I recommend doing a non-correcting check (so if anything went wrong you can repeat the rebuild). © Now you can pre-clear the drive with the pending sector ... this should result in that sector being reallocated -- or possibly even cleared altogether. Then you can add the drive to your array if the pending is cleared. Link to comment
trurl Posted September 13, 2016 Share Posted September 13, 2016 Stopping, unassigning old, starting to see missing, stopping, assigning new, not really necessary. Just shutdown, replace old disk with new disk, boot up, then assign the new disk to the old disks slot and start will rebuild. Link to comment
ehauk Posted September 15, 2016 Author Share Posted September 15, 2016 So I did the rebuild with the new disk and everything seems back to normal, but it did come back with these 3 messages. Should I be concerned? Event: unRAID Status Subject: Notice [TOWER] - array health report [FAIL] Description: Array has 5 disks (including parity) Importance: alert Parity - ST4000DM000-1F2168_Z302SH7E (sdg) - active 34 C [OK] Disk 1 - HGST_HMS5C4040ALE640_PL2331LAG6008J (sdf) - active 37 C [DISK INVALID] Disk 2 - Hitachi_HDS5C3020ALA632_ML0220F31JRX3N (sdd) - standby [OK] Disk 3 - WDC_WD20EZRX-00DC0B0_WD-WMC300847032 (sdb) - active 31 C [OK] Disk 4 - Hitachi_HDS5C3020ALA632_ML2220F30XHGSE (sde) - standby [OK] Data rebuild in progress. Total size: 4 TB Elapsed time: 10 hours, 16 minutes Current position: 3.38 TB (84.4 %) Estimated speed: 86.1 MB/sec Estimated finish: 2 hours, 1 minute Sync errors corrected: 1 Event: unRAID Disk 1 message Subject: Notice [TOWER] - Disk 1 returned to normal operation Description: HGST_HMS5C4040ALE640_PL2331LAG6008J (sdf) Importance: normal Event: unRAID Data rebuild: Subject: Notice [TOWER] - Data rebuild: finished (1 errors) Description: Duration: unavailable (no parity-check entries logged) Importance: warning Link to comment
JorgeB Posted September 15, 2016 Share Posted September 15, 2016 Subject: Notice [TOWER] - Data rebuild: finished (1 errors) This is normal if you are on v6.1.9, it's a bug. Link to comment
ehauk Posted September 15, 2016 Author Share Posted September 15, 2016 I am on v6.1.7. Is this still applicable? Link to comment
JorgeB Posted September 15, 2016 Share Posted September 15, 2016 Yes, I meant v6.1.9 or older, it was only fixed on v6.2. Link to comment
Recommended Posts
Archived
This topic is now archived and is closed to further replies.