unRAID4.7

 

I have just replaced the parity drive as it was showing high reallocated sectors, I used the guide below to replace it, I did run 3 counts of pre-clear on the replacement and it was good.

 

1. Run preclear on the drive a could of times (not necessary but will stress test the drive to find early failure)

2. Take out the old party by disconnecting it

3. Hook up new parity drive and assign it as parity once the computer is started back up

4. Let parity rebuild and then run a check on it to make sure nothing is going funky.

 

When it booted up the new drive was automatically allocated as parity, I did not have to assign it, is this normal? I did install the new parity into the same slot as the old one, so assume this is way it works if you place the new drive in the same slot.

 

The array status is "Stopped, updating parity" and the parity drive has a blue ball - does this mean all is good at the moment?

 

 

Sounds good to me.

Great, I'll just leave it alone until it's finished.

 

Thanks 

Great, I'll just leave it alone until it's finished.

 

Thanks 

It's never going to finish if you don't actually start the operation. There should be a check box saying something about yes I want to do this which when checked allows you to start the process.

Yeah, got it thanks. Slightly misleading, "Parity updating", "Parity sync pending" might have been better wording. Guess I need to update to the newer version some when.

 

Hmm, I did the start array and start sync and it now says "DISK_DSBL"...so its now disabled the parity disk.

 

What should I do now?

Yeah, got it thanks. Slightly misleading, "Parity updating", "Parity sync pending" might have been better wording. Guess I need to update to the newer version some when.

 

Hmm, I did the start array and start sync and it now says "DISK_DSBL"...so its now disabled the parity disk.

 

What should I do now?

Capture a full syslog from the current session before you reboot and zip it up and attach it here.

Well I think I screwed up saving the log, I did cat syslog > syslog.save but it wasn't there after rebooting.  Any way after rebooting it says parity is rebuilding, but the parity drive has a red ball and disk 4 is unformatted.

 

I think that after the first reboot, the one just after installing the new parity drive, that disk 4 may have been missing, after the second reboot it appeared again but unRAID thinks it is unformatted.

 

I'm going to leave for tonight and look at tomorrow night again.

You probably bumped a cable when changing disks. Shut down and check all power and sata cable connections to all drives. Then start it back up and report back. If disk4 is still unformatted then we will need to consider how to proceed.

I was thinking it might be something like that, it's 81% throuhg the parity check, I'll wait until that completes and then check SATA cables.

 

I still have the old parity drive, if it comes down to it, as I'm now wondering if I have lost data, can I use that to restore from?

 

 

I was thinking it might be something like that, it's 81% throuhg the parity check, I'll wait until that completes and then check SATA cables.

 

I still have the old parity drive, if it comes down to it, as I'm now wondering if I have lost data, can I use that to restore from?

Has anything attempted to write to the server since you started this?

A parity drive is only useful in combination with ALL other disks in the array (except one that you are trying to reconstruct). If the parity drive had been disconnected from the other drives and writes have occurred to the array during that time, the reconstruction will be negatively impacted in unpredictable ways.

Not AFAIK nothing written or read. I even unmapped drives on PCs that normally access it. After the first reboot the drives were not mounted, and parity drive had a blue ball.

 

Well the parity/rebuild completed and everything is green, except is still thinks disk4 is unformatted for some reason, and disk4 was not touched AFAIK.

 

I think I have lucked out here, I have my shares configured that all TV Shows and Movies go to disk1 or 4 at high water, and the bulk ended up on disk1.  I have lost a handful of movies that seemed to have been on disk4, and that seems about it.  I do have the important data backed up to an external drive and in the cloud so I'm not panicking.

 

So, what is the best way to proceed with the unformatted disk4? I'm not really sure what has happened here now!

 

Okay, I've just noticed something and not sure if it is relevant, disk4 is showing up as /dev/hda and not /dev/sdX, and I have a print out from a few years back and it was /dev/hda then as as well, so why is this? It has been working fine for nearly 4 years.

 

AHCI is enabled in the BIOS, as the other drives come up as sdX, and all drive are connected to the same controller, I will double check though.

 

Assuming nothing new has been added, can I format disk4, power down, add the old parity drive and restore everything that was on disk4?

Ok, the /dev/hda is an oddity with this motherboard, if I do CTRL A during boot and go into the AHCI BIOS only the first 4 drives plugged into SATA1,2,3 are displayed. When I go into the regular BIOS SATA ports 1,2,3,4 are showing unused but SATA port 5 is showing as disk4 and IDE.

 

I can only think that disk4 being unformatted is most likely down to me and finger trouble, but I would still like to know that if I format disk4 then add the old parity disk will recover the contents of disk4?

Ok, the /dev/hda is an oddity with this motherboard, if I do CTRL A during boot and go into the AHCI BIOS only the first 4 drives plugged into SATA1,2,3 are displayed. When I go into the regular BIOS SATA ports 1,2,3,4 are showing unused but SATA port 5 is showing as disk4 and IDE.

 

I can only think that disk4 being unformatted is most likely down to me and finger trouble, but I would still like to know that if I format disk4 then add the old parity disk will recover the contents of disk4?

DO NOT format disk4!

 

First of all, there would be no point if what you intend to do is rebuild it from the old parity. Formatting the drive will just create an empty file system on it. If you rebuild it from old parity, then ideally you will get a file system with all its files.

 

Formatting a drive is not a step you take to prepare a drive for use, it is a step you take when you want to create an empty file system. Many people seem to have some idea that formatting a drive is something you should always do before you try to use a disk.

 

There have been several cases where a user wanted to do a rebuild, but instead they formatted the drive. Formatting a drive actually writes to the drive, and unRAID updates parity to reflect the changes made to the disk when the empty file system is created. The end result is that you can't rebuild from parity then.

 

Instead of attempting to rebuild from the old parity, I think I would save that option for later, and instead first try to fix the filesystem.

The drive is not formatted, I tried running reiserfsck

 

Do you want to run this program?[N/Yes] (note need to type Yes if you do):Yes

reiserfs_open: the reiserfs superblock cannot be found on /dev/md4.
Failed to open the filesystem.

If the partition table has not been changed, and the partition is
valid  and  it really  contains  a reiserfs  partition,  then the
superblock  is corrupted and you need to run this utility with
--rebuild-sb.

 

 

 

Do you have backups of the files on disk4?

 

Do you have a spare drive you can rebuild to?

No backup for disk4, it was mainly TV shows, Movies etc, so I'm not desperate to get it back, it would nice if possible though. The more critical data on disk 2 & 3 is backed-up across multiple locations on-site and off-site, so I'm good there.

 

Just purchased 3 X 2TB WD RE4 WD2003FYYS-0 NOS $140 for the lot, so all the drives will be the same, so yes will have spare drives soon.

 

I have check all SATA cables etc, they are all good.

Just to recap

 

Unraid working fine,  noticed parity drive has growing pending sectors. As a preventive maintenance measure I decide to replace the drive with a new pre-cleared one.

 

I cleanly shutdown Unraid and install new parity drive into same slot. Old parity drive is disconnected.

 

I shutdown/turn off PCs with shares mounted to Unraid.

 

Reboot Unraid - new drive is automatically assigned as parity.  Array status is "stopped updating parity", parity drive has  blue ball, I cannot remember if it was blinking or not, all other drives are green, however, and this is where the issue may have started, I may not have noticed that disk4 may have got into a unformatted condition or not. I'm 100% sure that at this stage there were no red balled drives at all.

 

I post on forum asking if looks good, given the info I provided, it seems to be going ok so I leave it.

 

Several minutes pass and told that it will never finish the operation unless I start it, so I start the array/sync and then it immediately disables the parity drive "DISK_DSBL".

 

After checking cables, which look fine, I reboot, it says parity is rebuilding, and the parity drive has an orange ball I think (if that is possible), I may have been red. At this point I noticed that disk4 is unformatted. I leave to complete its parity rebuild, after which all green balls, but disk4 still show as unformatted.

 

I have not formatted disk4, as per suggestion, I tried reiserfsck and it seems not resolve the issue, I get the following

reiserfs_open: the reiserfs superblock cannot be found on /dev/md4.
Failed to open the filesystem.

 

I'm now wondering what to do next, I'm willing to risk using the old parity drive, if that is an option.

 

At this point I think it would be worth the risk to put the original set of drives back without drive 4, and tell unraid to use that set of drives to emulate drive 4 and see what is on the emulated version. To do that is a little complex, and it's very important to get the order of operations correct. I can take a stab at it, and possibly get it right, but I would be much more comfortable if you could get Tom M. from Limetech to walk you through it. Please try to email Tom and see if you can get some help, if not then one of us can possibly get it done.

 

Please post back to this thread with progress reports when you can get Tom involved so we know when you get it resolved one way or the other.

 

Tom usually responds to email pretty promptly, and that is his stated preference for direct support requests.

Where can I find contact info for Tom?  Is it tomb?

[email protected]

Where can I find contact info for Tom?  Is it tomb?

Email [email protected] and you should get a fairly prompt response. tomm@ is his direct email, but he watches the support email.

Everything you describe is correct for unRaid 4.7, and you followed the correct procedure (just remove old parity, install new parity in same slot - meaning same ide/sata port, and start parity sync).  The 'orange' ball is correct, just means data is invalid (which it is, because it's being created).

 

In unRaid-4 and previous, device assignment was done via 'slots' (PCI addresses) because in that version of linux kernel (and earlier) those never changed.  Starting with unRaid-5 that all changed and now devices are simply assigned using their ID (model/serial).

 

Back to this issue... The problem is something happened to disk4, meaning, as someone suggested, you bumped a cable in replacing your old parity disk.  Or something changed in your bios - if this is a 4-year-old server it's quite possible something caused a change in some bios setting.

 

The fact disk4 shows 'unformatted' really means that it can't be read correctly.  Probably if you were to post your system log it would be filled with numerous disk4 errors.

 

Probably what I would do is remove the parity drive again and reboot server without it.  Then find out why disk4 can't be read: check bios, cabling etc.  If you were unlucky enough that disk4 actually did decide to blow up during your parity upgrade, well you'll have to restore original parity, restore original config without disk4, and rebuild onto a new one.  But you are not at that point yet, because probably you have bad h/w you need to isolate first.  Don't write any new files since that will just make recovery harder.

Thanks for the reply.

 

Yes, I think something must have happened to disk4 when I opened the case.

 

I think something maybe note quite right with the motherboard.  As I mentioned earlier, it has 6 SATA ports, the first 4 connected in port number order are parity, disk1, disk2, disk3 and disk4, disk4 being on SATA port 3, this port is the one that comes as hdX and not sdX, so I guess the last 2 SATA ports must be on another onboard SATA controller.  Either way ACHI only picks up the first 4 SATA ports, the other 2 get picked-up in the regular BIOS fields, it's always been like this, and has worked fine for around 4 years and has been rebooted several times during that period and always worked.  I did double check the ACHI settings and the cables.  I will swap out disk4 SATA cable and see if it makes a difference, but I don't think it will.

 

If I mount disk4 via a Linux Rescue should I be able to see any files? And if there are no files there then I'm into the other options?

 

OK, I have found a setting in the BIOS, something like "Enable SATA IDE Mode", it was set to enable, if I set it to disable the AHCI BIOS then "see's" all the drives, set it back and it's back to seeing only the first 4 ports, and some point in the long and distant past in got enabled, and this is why disk4 was coming up a hdX and not sdX. Any I've tried various thing, reboot with out parity attached and disk4 on AHCI, unRAID then says there are 2 missing disks, rebooted again with parity disconnected, disk4 connected as SATA IDE mode and disk4 then appears again green ball but unformatted.

 

I think my last best shot is to try and use the old parity drive, if that doesn't work then I might as well go for a full system refresh, new motherboard, move up to unRAID6 OS and start over.

Well, I've found out that the keyboard is faulty, which hasn't been helping.