I had a disk fail, and replaced the disk with a precleared drive and started a parity rebuild.  this failed within 15 min.  here is the log from prior to replacing the failed drive.  

 

can I just re-start and try to repeat the rebuild?  the replacement disk was partially rebuilt with a small amount of data.

repository-syslog-20180407-1527.zip

Edited April 13, 20188 yr by kimifelipe

basic info on my server:  

 

Intel G620 processor

Supermicro C7P67 m/b, with two pcie x16 slots, 4 sata 3.0 and 4 sata 2.0 ports, dual realtek nics

Supermicro AOC-SASLP-MV8 8-Port SAS/SATA card

 

I built it in 2011

basically the question is, what happens when a rebuild fails?  do I need to re-clear the drive and start over?  the array knows the new drive now, but its data isn't rebuilt. 

 

Can anyone tell me basically what is going on?

There's no rebuild on the syslog you posted, there are several ATA errors and also disk corruption showing on disk3, next time post the complete diags.

 

As to your question, you can cancel and re-start a rebuild at any time.

Thank you for the reply.  I restarted the machine and tried to allow it to rebuild again, now I have two unmountable disks, one of them being the disk I am trying to replace - which is not in the same location as the disk I removed. 

 

is it possible this is a power supply problem?   attached is a log. 

 

thanks for any help. 

repository-syslog-20180408-0537.zip

3 hours ago, johnnie.black said:

next time post the complete diags.

Tools -> Diagnostics

I restarted the system and collected this.  do I need to bring the array back online and post diagnostics after that too?

repository-diagnostics-20180408-0628.zip

Disk9 dropped offline on the previous array start and that's why it turned unmountable, it should return to normal now, disk looks fine so it could be a connection issue or the controller it's connected to, since an SASLP is known to sometimes drop disks without a reason, check connections and try again.

 

Filesystem on Disk3 needs to be repaired like I posted above, you can do this before or after the rebuild, before would be best so you're not rebuilding an unfixable disk.

 

There's one unassigned 3TB Seagate also causing several ATA errors, if it's not in use disconnected it, if it is check connections.

thanks johnnie.  the unassigned Seagate is the drive that failed.  I had disk 3 precleared and ready as a hot spare.  I reassigned disk 3 to the new disk (now disk 3) and unassigned that Seagate. 

 

I have now pulled the Seagate, restarted, and am repairing the file system on the (partially rebuilt) new disk 3.  thanks for your help and i'll post when its complete and when the rebuild is complete.

Looks ok to me:  comments?  thank you!

 

ill read-only check consistency of the filesystem on /dev/md3
Will put log info to 'stdout'
###########
reiserfsck --check started at Sun Apr  8 08:58:47 2018
###########
Replaying journal:
Replaying journal: Done.
Reiserfs journal '/dev/md3' in blocks [18..8211]: 0 transactions replayed
Checking internal tree..  finished
Comparing bitmaps..finished
Checking Semantic tree:
finished
No corruptions found
There are on the filesystem:
Leaves 668460
Internal nodes 4094
Directories 4072
Other files 36534
Data block pointers 671967325 (0 of them are zero)
Safe links 0
###########
reiserfsck finished at Sun Apr  8 09:38:26 2018
###########

34 minutes ago, kimifelipe said:

Looks ok to me:  comments?

Yes.

16 hours ago, kimifelipe said:

basically the question is, what happens when a rebuild fails?  do I need to re-clear the drive and start over? 

Sounds like you have things under control.

 

For future reference, the only time unRAID requires a clear drive is when adding it to a new slot in an array with parity. The purpose of adding a clear drive to a new slot is so parity will remain valid. A clear disk is all zeros and so has no effect on parity. If you don't preclear a disk before adding it to a new slot, unRAID will clear it for you so parity will remain valid.

 

When rebuilding a disk like you did, it doesn't make any difference at all what is on the replacement disk since it will be completely overwritten by the rebuild. So, the incomplete rebuild you had the first try doesn't matter at all and you could just try again.

 

Some people also use preclear as a way to test a new disk even if they don't require a clear disk. 

so how do I restart the rebuild process?  bringing the array online doesn't trigger it.

1 hour ago, kimifelipe said:

I have now pulled the Seagate, restarted, and am repairing the file system on the (partially rebuilt) new disk 3.  thanks for your help and i'll post when its complete and when the rebuild is complete.

 

4 minutes ago, kimifelipe said:

so how do I restart the rebuild process?  bringing the array online doesn't trigger it.

 

This is a little confusing. Are you saying that the incomplete rebuild resulted in a disk that was no longer disabled? Or is the disk still disabled?

 

It is not actually possible to repair the filesystem on an incomplete rebuild. An incomplete rebuild is still a disabled disk. You must have been repairing the filesystem on the emulated disk3 instead.

 

the Seagate was disk 3.  I replaced it and started a parity rebuild before disconnecting the Seagate.  the new disk 3 is a new precleared disk.  the rebuild failed, and disk 3 is now emulated.  after pulling the Seagate out of the machine, it appears to be working properly, but I need to rebuild disk 3 ontothe new (current) disk 3.

OK, I get it. It thinks you need to replace the disabled disk before it will let you rebuild. That is completely normal.

 

Stop the array.

Unassign the disabled disk.

Start the array with no disk assigned to that slot. This will make it forget about the assigned disk.

Stop the array.

Reassign the disk to that slot, which will make it think it was replaced.

Start the array to restart the rebuild.

 

thanks tru.  I did that and the rebuild is in progress.  we are well past the point where it failed before removing that bad Seagate drive.  I believe it will be fine from here.

 

I have owned 6 or so of these ST3000DM's and have not had great success with them.  this is the second one that has failed in service, and I still have ST2000DM's humming along nicely in this server.  I have 3 more 3000's to replace, I am going to do them asap now. 

 

I will post a final update and a log after the rebuild is complete tomorrow.  I really appreciate t the help guys.

11 minutes ago, kimifelipe said:

I have owned 6 or so of these ST3000DM's and have not had great success with them.

 

They have a particularly bad reputation that has resulted in both a class action and a page on Wikipedia.

aw man I would love to have gotten my money back on these damn things. 

Update.  after having pulled the Seagate, I restarted the array and rebuild the drive onto a new disk.  that completed without errors.  I am now going to add and preclear a new drive, and work on completing the process of switching my array to xfs.  while I am doing that I will be replacing my remaining 3 ST3000DMs.

 

I have attached the syslog and diagnostics.

 

thank you again for your help.

repository-syslog-20180410-2132.zip

repository-diagnostics-20180410-2131.zip

Edited April 11, 20188 yr by kimifelipe

I finished preclearing a new disk and I see a lot of errors:  wrong csrf_token, over and over.  what does this mean?

 

the plugin reports that preclear has completed successfully.

 

log and diagnostics posted.  thanks for your help.

repository-diagnostics-20180412-0728.zip

repository-syslog-20180412-0724.zip

https://lime-technology.com/forums/topic/46802-faq-for-unraid-v6/?do=findComment&comment=545988

 

For future reference, there is no need to post diagnostics plus syslog, since diagnostics already includes syslog.

thanks for your replies, I understand both now.  my array seems to be running as well as ever now.  thanks.