Getting System Partition Failed on Raid Group 1 on multiple drives (SATA HDD) while verifying new disk to be added. Repair works but it happened twice already.

[undecided]

This is a new Xpenology server running 6.1.7 bare metal on an Intel 4790T with an LSI HBA ( LSI 9300-16i )and 6 6TB HDDs in RAID Group 1 (Raid-6,Btrfs). Most drives are WDC Red Plus with 2 Seagates. I also have a RAID Group 2 (Raid-6,Btrfs) which is 5 500GB SSDs. That has been stable. I had an addition EXOS drive set up as Hot Spare on Raid Group 1 and when the System Partition Failed message happened, that drive dropped out of being hot spare. The logs look suspicious:

2023/03/03 02:52:35  Write error at internal disk [12] sector 2153552.

2023/03/03 02:52:35  Write error at internal disk [8] sector 2153552.

2023/03/03 02:52:35 Write error at internal disk [7] sector 2153552.

2023/03/03 02:52:34 Write error at internal disk [13] sector 2153552.

2023/03/03 02:52:26 Write error at internal disk [14] sector 2153552.

How can all 5 drives fail on the same sector?

 

Thx

 

 

 

[Mary Andrea]

6 minutes ago, undecided said:

How can all 5 drives fail on the same sector?

Sometimes could be micro-cuts power. Have you tried change cable (Sata/Power) among disks? Check each connection through Board and Disks. Check plug tight. Try it. 

Edited March 3, 2023 by Mary Andrea

[undecided]

16 minutes ago, Mary Andrea said:

Sometimes could be micro-cuts power. Have you tried change cable (Sata/Power) among disks? Check each connection through Board and Disks. Check plug tight. Try it. 

Interesting. It was the same 5 disks both times. The weird thing is that they all failed on same sector. Now, my SATA power comes from different modular psu cables so I am pretty sure they do not share 1 cable and the one with the most plugs has 4 SATA power connectors on it and I have 5 drives fail simultaneously but I will double check.

[undecided]

Ooof, it just happened again. This time it was a different set of disks. System was running fine until I added this new disk and started the 'expansion' process which has been going for 2-3 days now. I wonder if the HBA is faulty.

[Mary Andrea]

4 minutes ago, undecided said:

I am pretty sure they do not share 1 cable and the on

Humm, share connection is not a problem, (its better for fluid electron vibrations) . Different factory quality, connectors, etc.etc. My solutions for micro-cut and permanently cache problems and slow sync, were minery splitters cables. -->  

[undecided]

Here's what I see

[undecided]

The Raid Group

[undecided]

There were no errors during regular operation of the RAID (copied 4-6TB to it). Only during this disk initialization process I see this problem.

[Mary Andrea]

Are you under DSM 6.x, really ? or 7.x ?? because in my experience  7.1 is more stable and faster to resync. -Maybe if you upgrade to 7.1 by easy way APRL . You could unplug groups/volumes, (save your usb loader 6.x), and with a new stick and one new HD install in the same server . After isntalation, if new DSM works ok, you plug connect one group/volume at time, and new DSM7 will recognize a migration, DSM7.1 will request you "if want to keep old configuration", so you have "to say OK", and old HDDs , in a no more 3 minutes charge HDs Volume into new DSM. After that you can charge by each group and no more 15' you rebuilt migration into 7.1 . I d'it and works, after that -in my case 14TB- depuration time was 8hs. (while in 6.3 were 2 days). So, its a new way.   

[undecided]

6.1.7 . I also found a post on STH about possibly needing to plug in the additional pciex power into the 9300-16i which I have not done as it is plugged into a pciex-16 port which should have enough power but who knows. I was very nervous going to 7.1. Been on 6.x for years.

[Mary Andrea]

 

[undecided]

Cool thanks. I plugged in additional power into the LSI 9300-16 by means of the 6-pin pciex cable from the PSU and everything works fine. It just needs extra power when stressed (like when adding disks). Thx