I just upgraded my parity drive 1TB WD1001FALS to a 2TB Samsung.  I did a parity check before the upgrade and that came up clean.  I upgraded the drive, and then did the initial parity sync.  I got 98 errors on disk 8, but everything completed, and it appears OK.  I noticed the sata cable on disk 8 was seated a little off to the side; I must have bumped it during the upgrade, but have since reseated it.  I don't understand how any errors could have been corrected since there was no parity data to fall back on.  Is my data safe?

 

What should i do next?  Do I need to revert back to my 1TB parity drive first?  Or should I do another parity check first w/ the new 2TB drive?

 

Here's my truncated syslog:

 

...
Mar 27 23:23:26 Server kernel: ata9: EH complete
Mar 27 23:23:30 Server kernel: ata9.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x0
Mar 27 23:23:30 Server kernel: ata9.00: irq_stat 0x00060002, device error via D2H FIS
Mar 27 23:23:30 Server kernel: ata9.00: failed command: READ DMA EXT
Mar 27 23:23:30 Server kernel: ata9.00: cmd 25/00:00:9f:f0:d5/00:04:4b:00:00/e0 tag 0 dma 524288 in
Mar 27 23:23:30 Server kernel:          res 51/40:ff:91:f1:d5/00:02:4b:00:00/e0 Emask 0x9 (media error)
Mar 27 23:23:30 Server kernel: ata9.00: status: { DRDY ERR }
Mar 27 23:23:30 Server kernel: ata9.00: error: { UNC }
Mar 27 23:23:30 Server kernel: ata9.00: configured for UDMA/100
Mar 27 23:23:30 Server kernel: sd 9:0:0:0: [sdj] Unhandled sense code
Mar 27 23:23:30 Server kernel: sd 9:0:0:0: [sdj] Result: hostbyte=0x00 driverbyte=0x08
Mar 27 23:23:30 Server kernel: sd 9:0:0:0: [sdj] Sense Key : 0x3 [current] [descriptor]
Mar 27 23:23:30 Server kernel: Descriptor sense data with sense descriptors (in hex):
Mar 27 23:23:30 Server kernel:         72 03 11 04 00 00 00 0c 00 0a 80 00 00 00 00 00 
Mar 27 23:23:30 Server kernel:         4b d5 f1 91 
Mar 27 23:23:30 Server kernel: sd 9:0:0:0: [sdj] ASC=0x11 ASCQ=0x4
Mar 27 23:23:30 Server kernel: sd 9:0:0:0: [sdj] CDB: cdb[0]=0x28: 28 00 4b d5 f0 9f 00 04 00 00
Mar 27 23:23:30 Server kernel: end_request: I/O error, dev sdj, sector 1272312209
Mar 27 23:23:30 Server kernel: ata9: EH complete
...
Mar 27 23:23:30 Server kernel: handle_stripe read error: 1272312328/5, count: 1
Mar 27 23:23:30 Server kernel: md: disk8 read error
...

 

And the entire thing:

syslog.txt

Well I realized I saved some new files to the server, so my old parity drive wouldn't be accurate anymore.  I went ahead and started a party check w/ the new 2TB drive.  Everything was green, so I assume there was nothing that unRAID couldn't figure out how to fix during the parity-sync.  Parity check speeds are the same as before at 55-58,000KB/s and I'm ~29% in w/ no errors so far.  I'll report back if I hit any more problems.  If those goes well, I'll move the old parity drive back into the array for storage.

Parity check completed w/ 98 errors again.  This time all the drives said 0 errors, but the parity check said, "(Last checked on 3/31/2010 2:33:34 AM, finding 98  errors.)".

 

All the lights are green.  Did unRAID fix the errors, or are there still problems I should address?

syslog.txt

Parity check completed w/ 98 errors again.  This time all the drives said 0 errors, but the parity check said, "(Last checked on 3/31/2010 2:33:34 AM, finding 98  errors.)".

 

All the lights are green.  Did unRAID fix the errors, or are there still problems I should address?

The parity "check" corrects the errors unless you used the command line to initiate it with a special NOCORRECT option. 

 

Therefore, you've already corrected the 98 errors that the last parity check found. 

 

Now, run another parity check.  It should find NO errors.  If it does, you have hardware issues to resolve.

The parity "check" corrects the errors unless you used the command line to initiate it with a special NOCORRECT option. 

 

Therefore, you've already corrected the 98 errors that the last parity check found. 

 

Now, run another parity check.  It should find NO errors.  If it does, you have hardware issues to resolve.

 

That's how I understand it as well.  What bothers me is I also had 98 errors during the parity "sync" when I first upgraded the parity drive.  I'm not sure how any errors could be corrected during that unless it just kept re-reading until it got the correct data.

 

Here's what I did:

 

1. Performed parity check w/ existing drives; no errors

2. Upgraded parity drive

3. Performed parity sync w/ 98 errors on disk 8

4. Reseated sata cable on disk 8

5. Performed parity check w/ 98 errors; no errors on any particular drive, though

 

 

I'll try another parity check and see what happens.

 

Thanks for your help so far!

(Last checked on 4/1/2010 3:00:27 AM, finding 0  errors.)

 

It looks like I'm good!