Today I came to machine to find half my array seemed to be missing and when I investigated the drive was fine and present but my Cache drive seems to have shat itself. Only a month or two old and supposedly a super long endurance drive but clearly NOT

 

I am seeing BTRFS errors all over the place and can't really use system inc my VMs which is largely how I use computer so reduced to laptop access.

 

I'd like to recover whatever data I can (I was on one cache drive and literally planning to order a duplicate drive today but not that brand now!) as I of course don have backups of the data. I can replace media but my appdata is all there.

 

Can anyone offer suggestions how to fix whatever can be fixed before I order TWO new drives for cache??  Diags attached

 

package.zip

Looks like the device dropped and came online again, reboot and post new diags after array start.

53 minutes ago, JorgeB said:

Looks like the device dropped and came online again, reboot and post new diags after array start.

 

Those diags ARE after a reboot and after I got it to reappear....  It looks to me (I dont know) that it needs some btrfs repair but I read that can make matters worse.. Is the SSD toast even if I can get my data back?

5 minutes ago, methanoid said:

Those diags ARE after a reboot and after

 

Them it dropped again, suggest swapping slots with the other device and see were the problem follows, could be failing.

I already moved it from a motherboard NVME slot to a PCIE x4 card with NVME slot - that was how I got it to "reappear"....  

 

Its 3m old so I'm gonna return for refund cos I dont trust this BUT... is there anything I can do to get my data back? btfrs repair or similar?

 

BTW can I also not use BTRFS for cache in future... I'm wondering what relative benefits are compared to XFS or EXT4 or other FS. Can you advise, please?

 

And I should say THANK YOU for replying. I appreciate you're not paid to

It shows as connected in the Cache slot but also appears in the Unassigned Devices but my /mnt/cache seems to have the docker image and precious little else implying its NOT mounted. IF I could recover data I could move that to another SSD, reformat this and test or just return it?

OK while I wait for help I have tried what I can to see what I can "fix" and if I unmount everything and mount the "faulty" drive with different parameters ALL my data seems to be there and intact

 

The logs report 

 

mkdir -p /mnt/cache

mount -t btrfs -o noatime,space_cache=v2 /dev/nvme1n1p1 /mnt/cache

 

But ls/dev shows nvme0 and nvme2 cos there is no nvme1 so I mounted

 

mount -t btrfs -o noatime,space_cache=v2 /dev/nvme2n2p1 /mnt/cache 

 

And all my files are there... its like the unRAID config is the only thing thats borked... so... how do I safely tell unRAID to use the "correct" address and get the system back up and running. It looks to me like the NVME may be okay? But worrying how/why it could just drop out... but I read all sorts of things about how "random" btrfs is compared to ext4.

 

Advice PLEASE someone? I've tried but I dont want to muck it up cos I am impatient with all the waiting for someone to assist cos LT relies on the community to support its product

 

 

So, New Config and preserve all disk slots and cache slots and everything seems to be working....  (Actually it didnt but once I swapped the Cache and VM drives over it seemed to kick it into realising that nvme1n1p1 was NOT in system and was now called nvme2n2p1.

 

But I am worried... is the NVME dying (nothing in SMART reports indicates), is BTRFS stable or even suitable for cache drive, Would it make sense to buy a 2nd identical SSD and mirror them?

10 hours ago, methanoid said:

is the NVME dying

Possibly, especially if you have already tried a different m.2 slot.

 

10 hours ago, methanoid said:

is BTRFS stable or even suitable for cache drive,

Yes.

 

10 hours ago, methanoid said:

Would it make sense to buy a 2nd identical SSD and mirror them?

I would not buy another one from that brand, never heard of it.