Just purchased 8 locking SATA cables to connect the 8 motherboard ports to the drives,

 

Just be aware of this if you use this for WD drives.  Not all locking cables have the "bump" that some WD drives require

 

http://lime-technology.com/forum/index.php?topic=36065.msg335979#msg335979

Just purchased 8 locking SATA cables to connect the 8 motherboard ports to the drives,

 

Just be aware of this if you use this for WD drives.  Not all locking cables have the "bump" that some WD drives require

 

http://lime-technology.com/forum/index.php?topic=36065.msg335979#msg335979

 

Thanks for pointing that out, all the hard drives are in hotswap Xcase caddies (rebranded Norco SS-500s), so I should not be affected by it.

If I stop the array, use Utils -> New config and reassign the drives, saying that parity is valid, do I lose the capability of rebuilding the drive which has been unmounted?

Well, I created a new config anyway, good news is that I don't have any read errors.

Bad news is that Total size: 4 TB 

Elapsed time: 18 hours, 59 minutes 

Current position: 119 GB (3.0 %) 

Estimated speed: 51.5 KB/sec 

Estimated finish: 862 days, 8 hours, 38 minutes 

Sync errors corrected: 794113 

 

I have reseated the cables, still waiting for the locking ones.

Well, I created a new config anyway, good news is that I don't have any read errors.

Bad news is that Total size: 4 TB 

Elapsed time: 18 hours, 59 minutes 

Current position: 119 GB (3.0 %) 

Estimated speed: 51.5 KB/sec 

Estimated finish: 862 days, 8 hours, 38 minutes 

Sync errors corrected: 794113 

 

I have reseated the cables, still waiting for the locking ones.

If you are getting lots of sync errors then this implies that in fact the parity was NOT valid and you should not have ticked the trust parity box.  Running a parity check which has to fix lots of parity errors would be significantly slower than simply running a parity sync to create new parity from the existing data disks.

Let's put that to the test. I will do a new config, new parity.

New config, and removed all sata cables, replaced them with shorter runs (only one that I found laying around is a latching one).

So far, much better results

Total size: 4 TB 

Elapsed time: 9 hours, 43 minutes 

Current position: 2.42 TB (60.6 %) 

Estimated speed: 123.6 MB/sec 

Estimated finish:  3 hours, 46 minutes

No read errors so far. Latching cables to arrive in a couple of days.

 

 

Once parity calculation is complete, I would like to compare the drive I replaced with the drive that was rebuilt, is there any way to do that without having to add the drive to the array, which would require for me to add parity? Perhaps add it as a cache drive or any other options which would not force a parity check?

Thanks

 

I was thinking of using rsynch like I used before

rsync --progress -avh /mnt/disk?/ /mnt/disk8/

 

unRAID Parity sync:: 02-12-2015 03:26

 

Notice [TOWER] - Parity sync: finished (0 errors)

Duration: 14 hours, 27 minutes, 33 seconds

Average speed: 76.9 MB/sec

 

Looks like cabling issues were stopping me from completing the parity calculations, hopefully they will be set straight by using sata latching cables.

I connected 8 drives directly to the motherboard, and 4 to the SAS2LP-MV8. Next, adding the old drive as cache and comparing it to the rebuilt drive.

Then migrating the data from the 5 remaining .5 TB disks across the drives with free space, removing them from the array and rebuilding parity.

Tried using it as cache, but it says the drive is unmountable unless I format the drive. Any other way for me to compare those two disks than creating a new configuration, which would then trigger a new parity calculation (another 18 hours of server chugging away)?

The old disk should mount on the cache slot as long as the cache file system type is set to auto or the correct fs of the old disk, you can check it by clicking on the cache disk.

I used the Unassigned Devices plugin to mount the drive.

Then I realised that the rebuilt drive was read only? So reiserfsck shows 136 found corruptions. A patter is afoot here.

I noticed that if I recalculate parity before I replace a drive, everything will go well. If I don't, the drives generated will be corrupted.

This doesn't bode well for the ultimate worse scenario, where a drive is truly dead (not disabled) so I can't recalculate parity before I replace a drive.

What is going on here? I thought that parity continues to update as you move files on to drives, but in my case something seems to be going wrong.