Jump to content

[SOLVED] Disk replacement problems


Recommended Posts

Hey guys

 

One of my drives went down today (its easily 5 years old so that's not really my big problem). I've obviously ordered a new drive to replace it and that's all fine, but I still want to be able to get to my stuff on the server while I await the new drive.

 

My set up is 2x 2TB, 2x 1TB and a 1.5TB, with a 2TB parity drive. I'm running 4.7 btw.

 

After a reboot, one of the 1TB drives wasn't being detected even in bios, and when I cleared the error and got into unraid, it generated a few errors for that device. After a few extra reboots and changes of cables to check its the device thats wrong I figure that yeah, the drives dead. So I pull the drive out, and reboot. No errors this time. The server will boot and I can get into the web interface, but it won't let me start the array. I've unassigned the missing the drive, but it still is showing up with a red ball and 'missing' on the main screen and telling me 'too many wrong or missing disks'.

 

Now, I'm cool with installing the new disk and rebuilding from parity...

 

But what I really want to know is if I can get the array to start without the missing disk (assuming that I can do that and still rebuild my parity later) or ideally just access the seperate drives ? Just about all my entertainment stuffs are on the server and it's going to be a LONG few days without anything to watch or listen to!

Link to comment

Hey guys

 

One of my drives went down today (its easily 5 years old so that's not really my big problem). I've obviously ordered a new drive to replace it and that's all fine, but I still want to be able to get to my stuff on the server while I await the new drive.

 

My set up is 2x 2TB, 2x 1TB and a 1.5TB, with a 2TB parity drive. I'm running 4.7 btw.

 

After a reboot, one of the 1TB drives wasn't being detected even in bios, and when I cleared the error and got into unraid, it generated a few errors for that device. After a few extra reboots and changes of cables to check its the device thats wrong I figure that yeah, the drives dead. So I pull the drive out, and reboot. No errors this time. The server will boot and I can get into the web interface, but it won't let me start the array. I've unassigned the missing the drive, but it still is showing up with a red ball and 'missing' on the main screen and telling me 'too many wrong or missing disks'.

 

Now, I'm cool with installing the new disk and rebuilding from parity...

 

But what I really want to know is if I can get the array to start without the missing disk (assuming that I can do that and still rebuild my parity later) or ideally just access the seperate drives ? Just about all my entertainment stuffs are on the server and it's going to be a LONG few days without anything to watch or listen to!

It sounds like you either have two failed disks, or you moved one to a new port and have a second that has failed.

 

Without a syslog, no other advice can be offered other than to put back the disks;/cables as they were and see if the array will start.

 

DO NOT set a new disk configuration, DO NOT format any disks, DO NOT attempt to re-calculate parity.  One mistake now and you potentially lose two disks worth of data.

 

You need to get back to where only one disk is failed, otherwise there is no way to use parity and the other disks to re-construct it onto its replacement.  (With only 1 failed disk you will be able to start the array)

 

Joe L.

Link to comment

I'm pretty certain it's just one disk that's failed.

 

The others are showing up properly in web management and such. It's that despite being physically removed and de-assigned, it's still telling me that I have a missing disk which is preventing the array from starting.

Link to comment

I'm pretty certain it's just one disk that's failed.

 

The others are showing up properly in web management and such. It's that despite being physically removed and de-assigned, it's still telling me that I have a missing disk which is preventing the array from starting.

I guess you did not see this in the prior post:

Without a syslog, no other advice can be offered other than to put back the disks;/cables as they were and see if the array will start.

Link to comment

During a disk issue, don't ever start re-assigning existing disks or un-assigning the failed drive.

 

You've screwed it up somehow, causing what appears to be 3 disks assigned improperly. Were you swapping cables between disks and then rebooting without looking at the interface screen to see you had messed up the other disk assignments? It appears that unRAID thinks the parity drive was just changed and that disk5 was already wrong from a previous boot. Also, every drive connected has been assigned a spot and there isn't a missing disk - there should be for example 2 lines for disk6 with the second line saying it's missing. So, where'd this failed disk go - disk6 maybe???

 

 

Oct 13 15:14:25 Tower kernel: md: import disk0: [8,80] (sdf) ST32000542AS     9XW058WS size: 1953514552
Oct 13 15:14:25 Tower kernel: md: disk0 replaced
Oct 13 15:14:25 Tower kernel: md: import disk1: [8,0] (sda) SAMSUNG HD204UI  S2H7J90B720039       size: 1953514552
Oct 13 15:14:25 Tower kernel: md: import disk2: [8,48] (sdd) SAMSUNG HD204UI  S2H7J1BZB04915       size: 1953514552
Oct 13 15:14:25 Tower kernel: md: import disk3: [8,64] (sde) SAMSUNG HD154UI  S1XWJ1BSC03939       size: 1465138552
Oct 13 15:14:25 Tower kernel: md: import disk4: [8,96] (sdg) WDC WD10EAVS-00D WD-WCAU41121487 size: 976762552
Oct 13 15:14:25 Tower kernel: md: import disk5: [8,16] (sdb) ST2000DL003-9VT1 6YD0VT3K size: 1953514552
Oct 13 15:14:25 Tower kernel: md: disk5 wrong

 

If you still know what disks below where then re-assign all the existing ones properly and see what happens.

 

Peter

 

Link to comment

My parity drive power cable got dislodged when I was putting the new disk in and I didn't notice until after a reboot, so that should account for it suddenly repapering.

 

According to the web interface all the drives are assigned properlly (ie the right name to the right slot) but disk 5. The disk named in the syslog as disk 5 is the new drive that has been assigned to the disk 5 slot, but the web interface is still red-balling that slot, giving me the new drive and the old drive name in in italics under it.

 

Btw, if it helps, I'm not hugely concerned about loss of data in this case, I just want to get my array started... there was practically nothing on the old disk 5 (it was the last added to the array and so nothing got transferred onto it), it was just added space.

 

 

Link to comment

I thought that might be the case, but after I telnet in and run initconfig, nothing changes. Devices are still assigned and I still can't start the array. Any thoughts?

 

Edit-

 

I've just poked around inside the flash share, and it doesn't seem to have created super.bak, and super.dat is still there unchanged. I figured maybe the stick was running as read only, but I can happily copy/move/delete files by hand so it wouldn't seem that is the case.

 

Can I manually re-name super.dat to super.bak ?

 

Edit again -

 

It would appear that yes, I can.

 

Parity is rebuilding and all my lovely data is there.

 

Thanks for the help guys, although I'm still curious as to why initconfig wasn't working. I'm thinking that this was the cause of my woes right form the beginning...

 

Also... yay for 2000minutes until parity is valid *sigh*.

Link to comment

Also... yay for 2000minutes until parity is valid *sigh*.

When I rebuild my parity it said something in the order of 1600 went to bed 8 hours later it said 18 minutes left. Initial calculation I guess doesn't include speed increase down the road. It is at best a rough estimate.

Link to comment

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...