DSM 6.2 RAID 1 volume crash after power outage. Need advice.

[Milamber]

Hi!

After moving my NAS to a different location, it turns out they do have power outages weekly...
As a result NAS doesn't work properly after forth power outage.

Setup:
DS918+ with DSM 6.2
4+4 GB RAM (upgraded with official Synology RAM)
No SSD cache
4x Seagate IronWolf 4TB (all formatted BTRFS)
Bay 1 and Bay 2 - RAID 1 (Volume 1)
Daily snapshots

After power outage NAS reported about Volume 1 crash (RAID 1). Other volumes are ok.
Initially NAS will power up, startup DSM and it was possible to view Folders on the crashed volume.
Later on it was not possible to view Folders on this volume and downloaded applications disappeared.
In the end NAS powers up, Status light blinks green with dead lights for all 4 disks...

There have been daily backups made for important data on Volume 1, so nothing serious has been lost so far.
Still I need to access folders on crashed volume and copy the data - my strategy is to access it via Ubuntu.

I'm in need of advice on how to proceed and dilemmas to solve.

1.
Is there any way to restore Volume1 and avoid all the hustle or somehow minimise it?

2.
If it is not possible to restore Volume1:
I will have to copy data from it, format "raid 1" drives, install them to NAS and install DSM. (If anyone has other ideas, please share)

Would I be able to access shared folders on 2 other drives on fresh DSM?

Should I continue using RAID 1 for redundancy or it is better to use second drive as a backup destination to avoid this hustle in the future?
RAID 1 volume is the only crashed...
Or completely different approach - SHR?

3.
Any advice on UPS to choose for DS918+


Please do share your experience.
I'll be very grateful for any advice
Thanks!

 

[jeyare]

NAS turned off
pull drives
one of Raid1 to SATA / USB or direct SATA to computer
Ubuntu is a good choice.
I have never tried to recover a marriage of devicemapper / btrfs = syno architecture. So don't even try 'testdisk', it only does ext (all) and FAT, exFAT, ntfs

I would try over lvm and ddrescue.
You will need a clean rescue disk
and a lot of patience.

 

[Rusty]

1.
Is there any way to restore Volume1 and avoid all the hustle or somehow minimise it?

2.
If it is not possible to restore Volume1:
I will have to copy data from it, format "raid 1" drives, install them to NAS and install DSM. (If anyone has other ideas, please share)

copy data, destroy the volume, recreate it, restore data

Should I continue using RAID 1 for redundancy or it is better to use second drive as a backup destination to avoid this hustle in the future?

having a backup as part of the device that holds the original data is not backup

RAID 1 volume is the only crashed...
Or completely different approach - SHR?

SHR is RAID1so in terms of redundancy you will not gain anything. In terms of expansion and drive density combination, you will have more options vs the conventional RAID.

Any advice on UPS to choose for DS918+

Almost any UPS will work just fine (EATON, CyberPower, APC...)

 

[Milamber]

NAS turned off
pull drives
one of Raid1 to SATA / USB or direct SATA to computer
Ubuntu is a good choice.
I have never tried to recover a marriage of devicemapper / btrfs = syno architecture. So don't even try 'testdisk', it only does ext (all) and FAT, exFAT, ntfs

I would try over lvm and ddrescue.
You will need a clean rescue disk
and a lot of patience.

Thanks for the answer!
I have just received a new Seagate IronWolf 4TB for recovery.
I'll try lvm and ddrescue (several runs) with one drive and will try to run nas with it, if it is ok - I'll insert another one and recreate RAID 1 with DSM.

copy data, destroy the volume, recreate it, restore data

Thanks for the answer!
There are several encrypted folders on the RAID1 drives, as well as on others without volumes being crashed. Would I be able to mount encrypted folders on fresh DSM if they recover?

having a backup as part of the device that holds the original data is not backup

Agreed. I never use backup within the same device. What I meant and did not clarify - was putting a drive into another NAS. I have DS115j collecting dust.
Since I don't see a point of having RAID1 with volume crashes.
Another idea id to use SSD as a main drive and HDD for storage (and backups to different destinations\devices).

SHR is RAID1so in terms of redundancy you will not gain anything. In terms of expansion and drive density combination, you will have more options vs the conventional RAID.

Considering I choose SHR for 2 or 4 drives, in case this one volume crashes I will have the same problem.
Do I understand it right?

Almost any UPS will work just fine (EATON, CyberPower, APC...)

Thanks!
I'm in a process of choosing between them

 

[Rusty]

Would I be able to mount encrypted folders on fresh DSM if they recover?

yes you will

Considering I choose SHR for 2 or 4 drives, in case this one volume crashes I will have the same problem.
Do I understand it right?

well she with 2 drives is mirror (raid1). shr with 4 drives is raid5 so not the same array setup.

 

[Milamber]

well she with 2 drives is mirror (raid1). shr with 4 drives is raid5 so not the same array setup.

I mean in terms of logical volumes if it makes sense. RAID5 \ SHR for 4 drives will "held" 1 logical volume and in case of checksum mismatch \ volume crash, I will end up with same issue and will have to recover and rebuild everything.
Basically I'm trying to understand which way to choose to avoid such problems in the future: RAID1 with 2 drives, RAID6 \ SHR with 4 drives or no raid at all. Backups are regular and UPS is on its way.
Considering the only crashed volume is on the drives with RAID1 and everything is fine with other drives and volumes without any RAID it makes choosing very unsettling since I have no prior experience with such failures.

 

[Rusty]

I mean in terms of logical volumes if it makes sense. RAID5 \ SHR for 4 drives will "held" 1 logical volume and in case of checksum mismatch \ volume crash, I will end up with same issue and will have to recover and rebuild everything.
Basically I'm trying to understand which way to choose to avoid such problems in the future: RAID1 with 2 drives, RAID6 \ SHR with 4 drives or no raid at all. Backups are regular and UPS is on its way.
Considering the only crashed volume is on the drives with RAID1 and everything is fine with other drives and volumes without any RAID it makes choosing very unsettling since I have no prior experience with such failures.

Think this was just an unfortunate situation. Personally I run multiple raid1,5,6 setups and haven’t noticed that one is preferred over the other in terms that they crash because of the actual array setup.

 

[Milamber]

After a long wait I have finally completed restoring my NAS to an operation mode and wanted to share my experience on the way in case anybody has similar issues.

NAS startup problem (Solved)
At one point my NAS didn't startup at all, fans started for 2 seconds then they stop and continue working. During that time no "light" from the status leds and no beeps. The mother board didn't want to startup. Tried several times.
Called several repair shops, and they confirmed "in 95% cases it either faulty mother board or power supply". I was unamused to say the least, but arranged a visit to the local repair shop.
For some reason I chose to open up NAS, clean it and look for a faulty part in the mother board (nothing sophisticated, just looking for "burn marks" with my eyes).
After cleaning, I decided to power on once again. For my surprise it did! A simple dust clean solved this problem. I must say, there was a lot of dust since it had been operating on the basement floor for several months until fateful power surge...

BTRFS volume repair \ Data recovery (somewhat solved)
Initially my plan was to use ddrescue to restore volume and get the data as suggested above. But my wait for a computer to leave on this task uninterrupted 24/7 had been extended, so I decided to go another way.
I bought a new disk and installed DSM on it. After that 2 disks with healthy volumes and without RAID were mounted. It worked like a charm - encrypted folders worked as well.
Once disk with crashed volume mounted, NAS started beeping informing about unhealthy volume. But it was possible to copy the data from unencrypted folders! So I did.
Next step was to repair a volume with Storage Manager in attempt to recover access to encrypted folders. It required a clean disk to re-establish RAID 1 for this crashed volume. Repairing took about a day.
After repair:

• DSM reported a volume as Healthy;
• all encrypted folders were still inaccessible - error during mount;
• copy data off some folders was very-very slow. 15 GB took one day...

In the end I had to erase repaired volume to avoid any problems like speed copy in the future. Obviously it wasn't repaired as expected.

Moral of the story—always use UPS and backup the data!
P.S. From now on my NAS has an overkill UPS and inaccessible data from encrypted folders was restored from backup