see this post, http://lime-technology.com/forum/index.php?topic=13866.0    I had a combination of things, the spin down issue and the loss of data issue. I used unmenu mount the old drives via usb and copied the data back.

 

 

I swapped my parity disk, and one smaller disk (disk 3) last week for larger drives and had no issues. I start the process with a third disk (disk 4) and now it has a bunch of read errors on disk 2 during the rebuilding.

 

Looking at some of the errors one of the other disks (seperate from the ones i replace) has current pending sectors and is starting to fail.

 

Did I screw up my data? is the parity not valid now?

 

Jul  5 11:28:59 Fileserver02 kernel: ata7.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x0
Jul  5 11:28:59 Fileserver02 kernel: ata7.00: irq_stat 0x40000001
Jul  5 11:28:59 Fileserver02 kernel: ata7.00: failed command: READ DMA EXT
Jul  5 11:28:59 Fileserver02 kernel: ata7.00: cmd 25/00:00:b7:74:80/00:04:53:00:00/e0 tag 0 dma 524288 in
Jul  5 11:28:59 Fileserver02 kernel:          res 51/40:00:f4:76:80/00:00:53:00:00/e0 Emask 0x9 (media error)
Jul  5 11:28:59 Fileserver02 kernel: ata7.00: status: { DRDY ERR }
Jul  5 11:28:59 Fileserver02 kernel: ata7.00: error: { UNC }
Jul  5 11:29:01 Fileserver02 kernel: ata7.00: configured for UDMA/133

syslog.zip

Disk2 is giving read errors. If the other disks are ok then you can try to rebuild disk2 on to itself.

yea, i will swap out disk 2 tonight. I'm wondering since it happened during an disk upgrade if it wrote bad data back to the parity

Parity should be fine. You can preclear disk2 and see if it passes. The disk may still be useable.

Before you get too carried away tell us what state your server is current in. Do not go ahead with any further changes until you evaluate the present state of the server.

 

The parity will still be OK but the new rebuilt disk2 will likely be bad since you were upgrading that disk while those read errors were occurring. unRAID must use a combination of all the other data disks and the parity disk to recreate a disk. The parity alone is worthless without all the other data disks so you must not think that having a good parity disk means your data is safe. The disk rebuild can't happen correctly if any one of the other disks is acting up and not reading correctly. Basically, there could be an error introduced on the rebuilt disk matching up with every time the bad disk gave a read error.

 

If things are bad with disk2 then it may be possible to re-install the old disk2 and use that combined with the parity and the other good disks to properly rebuild the failing disk.

 

So, answer these questions:

 

Is the data on the new disk2 all correct? Try reading or playing some/all of the files. Confirm the data as much as possible. Any errors here will cause errors when rebuilding the failing disk.

 

Do you still have the old disk2 as it was?

 

Have you written any new data to the server since upgrading disk2?

 

Peter

 

Before you get too carried away tell us what state your server is current in. Do not go ahead with any further changes until you evaluate the present state of the server.

 

The parity will still be OK but the new rebuilt disk2 will likely be bad since you were upgrading that disk while those read errors were occurring. unRAID must use a combination of all the other data disks and the parity disk to recreate a disk. The parity alone is worthless without all the other data disks so you must not think that having a good parity disk means your data is safe. The disk rebuild can't happen correctly if any one of the other disks is acting up and not reading correctly. Basically, there could be an error introduced on the rebuilt disk matching up with every time the bad disk gave a read error.

 

If things are bad with disk2 then it may be possible to re-install the old disk2 and use that combined with the parity and the other good disks to properly rebuild the failing disk.

 

So, answer these questions:

 

Is the data on the new disk2 all correct? Try reading or playing some/all of the files. Confirm the data as much as possible. Any errors here will cause errors when rebuilding the failing disk. I replaced parity drive went fine, I then upgraded disk 4 and during the rebuild of drive 4 that i started getting the errors from disk 2

 

Do you still have the old disk2 as it was? Yea I guess, based on the other posts I replaced drive 2 and it was almost done when I left it

 

Have you written any new data to the server since upgrading disk2? Yes, but its ok to loose it if i need to

 

Peter

 

hmm post keeps moving around, i posted it in Beta area, was i wrong?

I wrote my post wrong, I got confused and posted using disk2 to describe the one you had just upgraded. At any rate, you really didn't want to replace the failing disk2 until you had confirmed that the previous new disk (disk4 I believe) had been rebuilt correctly and was actually good. It sounds too late now so I hope it works out for you.

 

I expect some data damage considering the fact you were seeing sector errors on disk2 while disk4 was being rebuilt. Basically, if a sector on any other array disk can't be read then it's not possible to successfully rebuild the corresponding sector on the new disk. It's impossible to know what is damaged though so you'll just have to go looking for problems and hope you were lucky.

 

Peter

 

Well i still have disk 4 and 2 (originals). I should be able to hook these up via usb and copy the data back over no?

odd. sync finished, no drive errors still says it fixed 496 errors.

syslog-2011-07-07.zip