Hi all,

 

I currently have an unRAID v5.0.5 system that's been running on virtually auto-pilot for several years. It runs a monthly parity check and has always indicated no issue. The logs below occurred during this months parity check and it still showed up with 0 parity errors.

 

Setup is:

• Hardware: HP ProLiant MicroServer N36L 
• Data: 4 x 2TB WD20EARS
• Parity: 2TB WD20EARS

 

Have recently noticed the following:

Nov  1 05:14:42 storage kernel: ata1.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x0 (Errors)
Nov  1 05:14:42 storage kernel: ata1.00: irq_stat 0x40000001 (Drive related)
Nov  1 05:14:42 storage kernel: ata1.00: failed command: READ DMA EXT (Minor Issues)
Nov  1 05:14:42 storage kernel: ata1.00: cmd 25/00:00:68:62:72/00:04:b7:00:00/e0 tag 0 dma 524288 in (Drive related)
Nov  1 05:14:42 storage kernel:          res 51/40:9f:c8:64:72/00:01:b7:00:00/e0 Emask 0x9 (media error) (Errors)
Nov  1 05:14:42 storage kernel: ata1.00: status: { DRDY ERR } (Drive related)
Nov  1 05:14:42 storage kernel: ata1.00: error: { UNC } (Errors)
Nov  1 05:14:42 storage kernel: ata1.00: configured for UDMA/133 (Drive related)
Nov  1 05:14:42 storage kernel: sd 1:0:0:0: [sdb] Unhandled sense code (Drive related)
Nov  1 05:14:42 storage kernel: sd 1:0:0:0: [sdb]   (Drive related)
Nov  1 05:14:42 storage kernel: Result: hostbyte=0x00 driverbyte=0x08 (System)
Nov  1 05:14:42 storage kernel: sd 1:0:0:0: [sdb]   (Drive related)
Nov  1 05:14:42 storage kernel: Sense Key : 0x3 [current] [descriptor]
Nov  1 05:14:42 storage kernel: Descriptor sense data with sense descriptors (in hex):
Nov  1 05:14:42 storage kernel:         72 03 11 04 00 00 00 0c 00 0a 80 00 00 00 00 00 
Nov  1 05:14:42 storage kernel:         b7 72 64 c8 
Nov  1 05:14:42 storage kernel: sd 1:0:0:0: [sdb]   (Drive related)
Nov  1 05:14:42 storage kernel: ASC=0x11 ASCQ=0x4
Nov  1 05:14:42 storage kernel: sd 1:0:0:0: [sdb] CDB:  (Drive related)
Nov  1 05:14:42 storage kernel: cdb[0]=0x28: 28 00 b7 72 62 68 00 04 00 00
Nov  1 05:14:42 storage kernel: end_request: I/O error, dev sdb, sector 3077727432 (Errors)
Nov  1 05:14:42 storage kernel: ata1: EH complete (Drive related)
Nov  1 05:14:42 storage kernel: md: disk1 read error, sector=3077727368 (Errors)
Nov  1 05:14:42 storage kernel: md: disk1 read error, sector=3077727376 (Errors)
Nov  1 05:14:42 storage kernel: md: disk1 read error, sector=3077727384 (Errors)
Nov  1 05:14:42 storage kernel: md: disk1 read error, sector=3077727392 (Errors)
Nov  1 05:14:42 storage kernel: md: disk1 read error, sector=3077727400 (Errors)
Nov  1 05:14:42 storage kernel: md: disk1 read error, sector=3077727408 (Errors)
Nov  1 05:14:42 storage kernel: md: disk1 read error, sector=3077727416 (Errors)

 

Disk 1 has about 1TB of data on it. All up the array is about 50% utilised.

 

I figure the disk above is on its way out, so a WD40EFRX 4TB RED drive is on its way.

 

What I'd like to do is:

• make the new 4TB the parity disk
• have the existing parity assume the role of the failing disk 1

 

Have found this link,.. the block titled "The procedure" seems relevent to me.
https://wiki.lime-technology.com/The_parity_swap_procedure

 

Every guide I've read always recommends preclearing, but have also read that its not necessary if its a parity drive ?

 

I'm figuring the following is what i need to do:

1. Given I have no spare SATA spots in my system, I follow steps 1 to 8.
2. New drive will then be unassigned.
3. Run the pre-clear script against the new drive ? If its needed ?
4. Continue on from step 9

 

So my queries:

• is preclear needed if its role will be that of a parity drive ?
• am I following the right guide/steps for what I want to do ?
• am I missing anything ?

 

Any guidance would be appreciated.

 

Thanks in advance,

 

Jim.....

 

smart.txt

Edited November 3, 20178 yr by jimbo123
attached log

1 hour ago, jimbo123 said:

is preclear needed if its role will be that of a parity drive ?

 

preclearing in this case would only be needed if you wanted to test the disk to make sure it's good before using it in the array, it's not necessary for the procedure itself.

 

1 hour ago, jimbo123 said:

am I following the right guide/steps for what I want to do ?

 

yes

Drive arrived during the week,.. new drive is now going through a preclear,.. expect it should finish overnight and I can then start the configuring when I get home from work tomorrow.

 

So the situation I'm in is:

• Went through steps 1 to 8
• Array is running minus the the drive I pulled.
• New drive is preclearing in its slot.

I assume that when the preclear finishes, that I:

8b. stop the array

9. unassign the parity disk (can I do this without losing any data ? as the old disk is not replaced yet and I'll now also be a parity drive out ?)

10. assign the new drive as the parity

11. assign old parity drive into the array

12, 13 , 14 as per guide.

 

Step 9 is what I'm unsure of ? 

 

Or should I add:

8c. assign the new drive into the array (taking the slot of the failing drive I removed) - let it rebuild and sort itself out

 

and then continue with step 9 and onwards ?

 

Jim.....

5 minutes ago, jimbo123 said:

9. unassign the parity disk (can I do this without losing any data ? as the old disk is not replaced yet and I'll now also be a parity drive out ?)

 

Yes, that's part of the procedure, just follow it.

 

6 minutes ago, jimbo123 said:

Or should I add:

8c. assign the new drive into the array (taking the slot of the failing drive I removed) - let it rebuild and sort itself out

 

Isn't the new disk larger than parity and why you're doing the parity swap?

Yes, new disk is larger,.. and yes I want it as parity as it then gives me the ability to upgrade the others as I need to for additional storage.

 

I re-read steps 12 and 13 and its now more clearer to me. Thanks JB.

 

Just needed some reassurance