[HELP] 2-3 drive fail + 1 drive read error

[Xili]

Hello

The recovery of my important files is completed without any particular worries.
I launched the parity check, obviously a lot of error in the parity, I suppose because of the checks started during reading problems. I imagine that it is the rewrites on the disc which make the check is slower than normal 80mb/s against 120 in normal time

[JorgeB]

So you are using the old parity drives? And yes it's normal for the speed to be slower when it's correcting errors. 

[Xili]

yes I took over the old parity disks first.
yes it does not every me, it's a lot of rewriting to do. And I haven't restarted any service on the side to limit disk usage so that it's as fast as possible. This should end Thursday morning.
I will get back to you at that time.

[Xili]

hello

the parity check finally finished earlier, it has accelerated well beyond 6to, 180mb/s, because none of the disks are larger than this size by those of partition for the moment

Last check completed on Wed 09 Nov 2022 02:43:02 PM CET (today)
Duration: 1 day, 14 hours, 5 minutes, 54 seconds. Average speed: 102.1 MB/s
Finding 1465130620 errors

I have not yet restarted my containers, but a priori nothing prevents it, everything seems ok to me?

I am posting my diagnostics here. Is there anything shocking in it?

 

Small question: I have 2 disks of 16to to inject and put in parity since they are the biggest, to pass the 14to in the array. What is the best way to avoid taking risks? I imagine that it is better to avoid changing the two at once, but rather one by one. But during this change can I immediately integrate the 14 into the array instead of another, or is it better to do it in two steps?

tower-diagnostics-20221109-1747.zip

[JorgeB]

No disk errors during the sync so that is good but since you are using a Ryzen server with overclocked RAM and that is known to sometimes cause data corruption I would recommend doing a non correct parity check to confirm before doing anything else, and change RAM speeds to max supported.

 

 

[Xili]

a non correct parity check ? I'm not sure what it is?

 

 

ok so for the ram I am with 4 bars with a 2700, so I put them at 1866, correct?

[JorgeB]

5 minutes ago, Xili said:

a non correct parity check ? I'm not sure what it is?

Just start a parity check with "write corrections to parity" unchecked, but it it detects errors stop and run a correcting check.

 

6 minutes ago, Xili said:

ok so for the ram I am with 4 bars with a 2700, so I put them at 1866, correct?

Correct, should be set @ 1866MT/s with 4 dual rank DIMMs

 

[Xili]

after research in the bios, and in the doc next door, I can't find a way to go below 2667

 

edit : or do I have to go to 3733 in dual, to do 1866 on each?

Edited November 9, 2022 by Xili

[JorgeB]

Should be 1866, make sure XMP is disabled. 

[Xili]

Hello
Sorry for the response time, it's been a busy week....

I did the checks indicated above and ended up restarting everything + moving the important files to a single disk, so I know if a disk is critical or not... , and I will have to quickly put an external backup in square .
For the ram, I let you see my screenshots of the bios, impossible to manually modify the frequency, I can only go down to 2667 as I have with the xmp deactivated

 

Edited November 20, 2022 by Xili

[JorgeB]

You'll need to consult the board manual or contact Gigabyte.