Hi, Hoping someone can advise me as I am stuck after a couple of problems with a 2 disk add to my array. (5.0 main release, Pro license)

 

All was well but I was running low on space so added 2x 4TB disks to existing 4TB parity, 2x 4TB and 1x 2TB data disks.

 

Both disks did zeroizing simultaneously without issues.

Short description is that both times the format button was available, the web GUI stopped responding after a short while.

On logging into CLI and issuing a powerdown each time, on resuming, that disk showed as not needing a format and then would join the array.

Of course, I lost parity sync through this so I initiated a parity sync and then I got massive error  quantities on all disks.

 

I believe this was a power problem with sharing 2 power busses across 4 bays of the host so I redistributed the disks in the chassis slots the errors stopped and the parity went through cleanly. Still original 3 data disks intact and 2 new disks reporting included in the available space.

 

Only rouble is I cannot write to either of these drives at all. Not in the share (SMB), not as individual disk mounts, not in the CLI, no go.

 

Not sure what my options are but I really don't want to lose the contents of the 3 initial data disks.

 

Anyone got any suggestions? I have no replacement disks to try and I think any rebuild would put me in the exact predicament I am in now.

 

Can I format an individual disk in the array somehow?

 

Is the restore array configuration destructive to the content of already formatted disks if I remove the two new disks and try adding them again?

(The manual is a little crypticy worded on this.)

 

Any advice appreciated.

 

fatpipe.

Did you do a pre-clear before adding the new disks to the array to check that they were OK.

Which model PSU?

Thanks dgaschk and itimpi for responding.

 

Some very usefull info in the additional part of your post dgaschk which may have helped me diagnose an plan a new course of action. Firstly though, answers to your questions.

 

PSU is Thermaltake TP-750M which boasts a single +12V rail of 60A. Trouble is it only came with two looms for periperal connector (traditional 4 pin) power with 3 connectors each. My chassis would need 13 such connectors to be connected fully. As I said, my suspicion with power was that it didn't like the two connectors being used across drive bay rows sharing on one controller. I passed through another controller (Super Micro AOC-SAS2LP-MV8) and moved the 4TB disks apart from parity disk over to it and all worked reliably after that except for the read only problem at the end. As you might have guessed, it is a VM in VMware ESXi 5.5

 

I did not pre-clear the drives. It did do the zeroizing without incident though. Also it seems it wrote valid reiserfs file systems on the 2 new drives according to reiserfsck so maybe they aren't the dead loss I thought they were.

 

Attaching the boot log. Near the end is an attempt to put data on the Disk4. Interesting the error looks like it is Disk2 that is write protected. This is not the case, I can write to disks 1, 2 and 3 still.

 

Also attaching the output of reiserfsck checks.

First I ran it on Disk1 to sanity check the results. All looks good there.

Ran on Disk4 and it reported and wrote some outstanding journal entries (2). Hmm.

Ran on Disk5 and it came up clean entirely.

 

I have put the array back online and blocked disk4 from the share and I am able to write to the other 4 disks (obviously mostly to disk5 because it is the most empty) so disk4 is the problem only now. Any suggestions to correct it?

Would this be all it needs?

from disk4 reiserfsck "2 found corruptions can be fixed when running with --fix-fixable"

 

 

Regards,

fatpipe

zStore_Log_File.txt

zStore_reiserfsck_D1_D4_and_D5.txt

Run reiserfsck with --fix-fixable.

disk4 might be set as read-only because of the corruptions that were discovered by reiserfsck. Run the command dgaschk mentioned.

OK. I am good now.

 

Used the --fix-fixable options and disk 4 (which I had blocked out by excluding it from the share) is now read-write and working fine.

 

Thanks all for your replies and help.

 

Thanks.

fatpipe