Parity drive disabled

[trurl]

29 minutes ago, trurl said:

cloning parity and disk4 before doing anything else, but that would mean even more spare disks.

Just noticed a 14TB unassigned disk, so that could be used to clone parity (if it is empty). Presumably it will be replacing parity anyway.

[JonathanM]

Cloning disks with poor health is better handled by ddrescue instead of dd, it's purpose built to get every last bit of good data transferred to a new drive.

 

Problem is, this is going to be a lengthy process to get to a point where we even know IF we can get anything usable with parity emulation.

 

Earlier in the thread there was a sense of urgency about getting the server up and running regardless of data loss on some of the drives. Where are we on the balance of time vs value of all array drive data? Is there another PC available other than the server available to work on data rescue? Maybe pull all the current parity array drives out, add the 14TB as disk1, put the cache pool back, and work from there?

[banterer]

The 14TB has a lot of other backups on, some of which have been transferred to the array and some haven't, so I'd rather keep that aside for now. It was my intention to eventually use it for parity, so that I can put whichever size disks I like in the array.

 

I have a 3TB WD Red arriving tomorrow, so it's kind of sounding like maybe ddrescue from a manually mounted disk4 -> new disk, then replace that in the array and try to start it? 

[trurl]

3 hours ago, banterer said:

ddrescue from a manually mounted disk4 -> new disk, then replace that in the array and try to start it? 

I guess we can try that to start with. Let us know when the disk arrives and we can get into the specific details of how to proceed.

[banterer]

Ok, disk installed as sdg. Disk4 is showing as sdh. 

 

So having a readup on ddrescue the syntax I want according to this post is:
 

ddrescue -d -f /dev/sdh /dev/sdg clone.logfile

 

But what's the best way to install ddrescue? 

[JorgeB]

https://forums.unraid.net/topic/46802-faq-for-unraid-v6/?do=findComment&comment=525075

 

[banterer]

Ok so it's almost done. Is it then just a case of selecting the clone in the array settings on 'main'?

[trurl]

Sorry, ignore my deleted post. In addition to what I said, we need to get spare disk3 into the New Config along with the clone disk4.

 

Do you have a spare to assign as disk3?

[banterer]

6 minutes ago, trurl said:

Do you have a spare to assign as disk3?

Can I use the 'old' disk4?

[banterer]

(I appreciate it's not ideal, but I only have so much money I've spend way too much on hardware already...)

[banterer]

If I need to, I'll order another...

[trurl]

6 minutes ago, banterer said:

Can I use the 'old' disk4?

We are going to disable it immediately anyway, so you could do that and we can see if the rest of the array can emulate disk3. If we can get disk3 emulated and mountable, we could even copy its data.

 

Instead of doing all the instructions at once, let's take this a little at a time and check how things are progressing along the way.

 

Tools - New Config, Retain All, Apply.

 

Assign clone as disk4, assign former disk4 as disk3. Don't do anything else until further instructions.

 

Then post a screenshot of Main - Array Devices.

[banterer]

15 hours ago, trurl said:

...

Tools - New Config, Retain All, Apply.

 

Assign clone as disk4, assign former disk4 as disk3. Don't do anything else until further instructions.

 

Then post a screenshot of Main - Array Devices.

 

Argh, just realised it didn't complete, I think. I'm guessing my Mac went to sleep and it dropped the terminal connection. I guess that this would stop the process? And that the only way forward will be to do it again and disable all sleep / screensavers etc etc...

[JorgeB]

If you use the same log file (/boot/ddescue.log) it will resume.

[banterer]

Ok, finished. So the results of ddrescue look good to me?! 

ddrescue -f /dev/sdh /dev/sdg /boot/ddrescue.log
GNU ddrescue 1.26
Press Ctrl-C to interrupt
Initial status (read from mapfile)
rescued: 2544 GB, tried: 356352 B, bad-sector: 0 B, bad areas: 0

Current status
     ipos:    3000 GB, non-trimmed:   356352 B,  current rate:  95813 kB/s
     opos:    3000 GB, non-scraped:        0 B,  average rate:  41779 kB/s
non-tried:  300154 kB,  bad-sector:        0 B,    error rate:       0 B/s
  rescued:    3000 GB,   bad areas:        0,        run time:  3h  1m 48s
pct rescued:   99.98%, read errors:        0,  remaining time:          4s
                              time since last successful read:          0s
Copying non-tried blocks... Pass 1 (forwards)
     ipos:    1342 GB, non-trimmed:   667648 B,  current rate:     455 B/s
     opos:    1342 GB, non-scraped:        0 B,  average rate:  40229 kB/s
non-tried:   65732 kB,  bad-sector:        0 B,    error rate:    6826 B/s
  rescued:    3000 GB,   bad areas:        0,        run time:  3h  8m 56s
pct rescued:   99.99%, read errors:        6,  remaining time:         41s
                              time since last successful read:          0s
Copying non-tried blocks... Pass 2 (backwards)
     ipos:    1342 GB, non-trimmed:   823296 B,  current rate:    3218 B/s
     opos:    1342 GB, non-scraped:        0 B,  average rate:  39713 kB/s
non-tried:   46333 kB,  bad-sector:        0 B,    error rate:    1462 B/s
  rescued:    3000 GB,   bad areas:        0,        run time:  3h 11m 24s
pct rescued:   99.99%, read errors:       10,  remaining time:          3m
                              time since last successful read:          0s
Copying non-tried blocks... Pass 4 (backwards)
     ipos:    2108 GB, non-trimmed:    1998 kB,  current rate:   1048 kB/s
     opos:    2108 GB, non-scraped:        0 B,  average rate:  38653 kB/s
non-tried:        0 B,  bad-sector:        0 B,    error rate:       0 B/s
  rescued:    3000 GB,   bad areas:        0,        run time:  3h 16m 40s
pct rescued:   99.99%, read errors:       35,  remaining time:          5s
                              time since last successful read:          0s
Copying non-tried blocks... Pass 5 (forwards)
     ipos:    2109 GB, non-trimmed:        0 B,  current rate:   19114 B/s
     opos:    2109 GB, non-scraped:   785408 B,  average rate:  37667 kB/s
non-tried:        0 B,  bad-sector:    34304 B,    error rate:     170 B/s
  rescued:    3000 GB,   bad areas:       67,        run time:  3h 21m 49s
pct rescued:   99.99%, read errors:      102,  remaining time:      2m 11s
                              time since last successful read:          0s
Trimming failed blocks... (forwards)
     ipos:    2109 GB, non-trimmed:        0 B,  current rate:       0 B/s
     opos:    2109 GB, non-scraped:        0 B,  average rate:  29828 kB/s
non-tried:        0 B,  bad-sector:   466944 B,    error rate:       0 B/s
  rescued:    3000 GB,   bad areas:      100,        run time:  4h 14m 50s
pct rescued:   99.99%, read errors:      947,  remaining time:          0s
                              time since last successful read:         15s
Scraping failed blocks... (forwards)
Finished

 

And here's my 'main' tab. with the old disk4 (sdh) in slot 3, and the new cloned disk4 (sdg) in slot 4

 

[trurl]

I don't seem to have any diagnostics with that clone in it, but all other assignments look like what we want so I assume that one is OK too.

 

Going forward, I will refer to these disks according to their new assignments.

 

In Main - Array Operation, check BOTH parity valid and maintenance mode boxes, then start the array. This will accept all the disks as assigned into a new array without actually trying to access or change any of the disk contents.

 

Then, stop the array, unassign disk3, start the array with disk3 unassigned. This will disable disk3 and emulate it from the rest of the array.

 

Then, post a screenshot of Main - Array Devices and new diagnostics.

[banterer]

All done so far...

 

tower-diagnostics-20230219-1100.zip

[JorgeB]

Check filesystem on the emulated disk3.

[banterer]

4 hours ago, JorgeB said:

Check filesystem on the emulated disk3.

It’s hung on stopping the array 😟

 

reboot?

[trurl]

post new diagnostics

[banterer]

3 hours ago, trurl said:

post new diagnostics

tower-diagnostics-20230219-1934.zip

[trurl]

You really should disable Docker and VM Manager in Settings until you get your array stable again.

 

Your bad parity disk is causing problems of its own, and it needs to cooperate in order to make any progress with disk3.

 

Maybe time to ddrescue that one too.

[banterer]

5 minutes ago, trurl said:

You really should disable Docker and VM Manager in Settings until you get your array stable again.

 

Your bad parity disk is causing problems of its own, and it needs to cooperate in order to make any progress with disk3.

 

Maybe time to ddrescue that one too.

Ugh, really the best option? If so I'll have to order another one. The irony being I got unraid so I could make use of all the different size disks I have, but I keep running out of 3TB ones 

 

 

So two questions I guess:

1. Is there any way to use a different size disk than 3TB to ddrescue parity onto?

2. I've disabled docker and VMs, can we get any idea whether it's worth buying another drive (ie the pros and cons of sticking with parity as is, or not)?

 

[trurl]

You can't use a data disk larger than parity, but, usually, you can use a larger disk for parity. Not entirely sure how that would play with ddrescue but it seems like it should work, since none of the larger part of parity is needed for emulating disk3.

 

Let's get another opinion on that though since I haven't done it.

@JonathanM

@JorgeB

 

[JonathanM]

Theoretically a larger parity drive would be fine, especially if you precleared it before the clone operation to ensure the remaining space was for sure zeroed.

 

However, this isn't something I've even come close to experimenting with, hopefully JorgeB with all his test arrays can say it will work for sure.