unRAID 4.4 Rebuild Failure

[jazbo8]

One of the hard disk (1TB) gave out, so I got a larger one (2TB), the system spotted the new disk (blue ball), I checked "I'm sure" button to start the rebuild process, but after a few minutes, the new drive showed a bunch of errors, and the system showed a red ball next to it. I have done two rebuilds in the past, both time, it worked without problem, but not sure what's happening this time around, any suggestion will be much appreciated.

[trurl]

 

[SSD]

Log would be helpful, but I expect your swapping out of old disk and swapping in of the new disk disrupted the delicate cabling.

[jazbo8]

Thanks for the replies, I tried another disk with the same result, after re-building the array, the new drive still shows a red ball. As you can see in the syslog, disk6 or /dev/sdi is the one with the problem. The data appears to be ok, I can see and access the files, how to clear the red ball and secure the array...

 

BTW, I am still on 4.4.2

syslog.txt

Edited March 7, 2017 by jazbo8

[JorgeB]

Can you try without that port multiplier? If not at least swap with the free port.

[jazbo8]

Do you mean moving the drive to another slot/port? I do have another open/un-used one.

[JorgeB]

Yes, if you must use that port multiplier swap the problem disk to use the available port.

[jazbo8]

Ok, let me try and report back.

[jazbo8]

I moved the new drive to another port as suggested, assign it, the drive shows a blue ball, but then I get this message on the UI.

 

Invalid expansion      You may not add new disks when there exists missing, wrong, or disabled disk(s).

 

 

[JorgeB]

That error implies you're assigning it as a new disk instead of replacing the missing disk, check the slot you're assigning it to.

[itimpi]

4 hours ago, johnnie.black said:

That error implies you're assigning it as a new disk instead of replacing the missing disk, check the slot you're assigning it to.

I have a feeling that the older versions of unRAID identified disks by the port they were connected to rather than the disk serial number which means it is not that easy to simply move a disk to another position..

[JorgeB]

10 minutes ago, itimpi said:

I have a feeling that the older versions of unRAID identified disks by the port they were connected to rather than the disk serial number which means it is not that easy to simply move a disk to another position..

 

Forgot about that, I bet you're right.

[JorgeB]

Thinking a little more about it you probably can't change port because you already started rebuilding in the current slot, but this should work:

 

Connect disk in old port, start array, cancel rebuild, stop array, unassign disk, start array with that disk unassigned, stop array, shutdown and change disk port, power on, assign disk and it should start rebuilding.

 

If it doesn't work let me know so I can fire up my v4.7 server I keep mostly for nostalgia so I can check the correct procedure.

[jazbo8]

Ok, let me try that and see what happens...

[jazbo8]

Well, I decided to copy the bad disk data to an external drive since no matter what I tried, the system could not rebuild the disk. Anyway, once all the copying was done, I used the Restore button to shrink the array, but the system hung in just a few minutes, so now the parity is all messed up. Thinking it might be a hardware or bad-wiring problem, I unplugged and re-plugged all the cables, made sure all the cards were properly seated, etc. During the troubleshooting, it appeared that the e-SATA card was bad, so I replaced it with another one.

 

Then I fired up the system, now all the drives showed up as new (blue), is this normal? I have two options on the screen - 1) start the array and rebuild the parity drive or 2) Restore (which I did already). What do you think is the best course of action? I'm worried that I might lost all my data if the system messes again during the rebuild...

syslog4.txt

[JorgeB]

Doing a restore will just make all disks new (blue), like they already are, you should start the array to begin the parity sync.

[jazbo8]

Ok, fingers crossed... Another question, in case the parity sync does not complete (as we have mains outage from time to time), will that harm the data on the other drives? The way I understand it, the parity sync reads from all the data disks, and writes to the parity disk only, so all the data on the disks are preserved, correct?

[JorgeB]

Parity sync not completing is not a problem, power cuts can be, they can cause file system corruption.

 

If the sync doesn't complete due to an error grab the syslog before rebooting.

Edited March 23, 2017 by johnnie.black

[jazbo8]

I think I have something more seriously wrong with my system, the parity sync stops after just a minute or two and the whole system hangs. I suspect my motherboard might be bad... Will investigate further.

[jazbo8]

I decided to upgrade to V.6 - a fresh start, but before I do that, is it ok that the parity sync did not complete under V.4.4.2, i.e., the array is unprotected? I hope upon boot up, V6 will recognize all the drives and start the parity sync. Could someone please confirm this is ok?

 

TIA,

jaz

Edited March 29, 2017 by jazbo8

[JorgeB]

Upgrading from v4 will require a clean install, you'll need to do a new config, assign all disks (don't trust parity) and start the array to begin the parity sync, see here for more details:

 

http://lime-technology.com/wiki/index.php/Upgrading_to_UnRAID_v6

[jazbo8]

Yes, a clean install for sure. I read the above linked document several times, and it made no mention of the parity sync, whereas in an earlier document (upgrading from 4 to 5 I think), it says that the parity sync must be completed before the upgrade can take place, so I just want to make sure. Anyway, It seems that I am ready to upgrade to V.6, fingers crossed (again). Thanks for your help as always.

[JorgeB]

You'll need to re-assign all your disks, since parity is not valid don't check the "parity is already valid" checkbox next to the array start button, so on array start a parity sync will begin.