Possible to retrieve lost data?

[Keyakinan]

This was all months ago, and I just couldn't find the energy to keep trying to fix it. I try to remember as much information as I can and be as clear as possible, but as you can read below, I'm not too well versed in this and was just burned out.

I had 2 4tb drives in mirror. One drive suddenly failed, but that was fine (I thought) because it was in mirror.
I tried to remount the pool a few times, and sometimes it worked a bit, but I couldn't open anything, or it took forever, so I couldn't copy the data over and format the drives. I tried to find the non-broken drive and mounted it in a VM (Ubuntu I think) with also not great success, the pool was there but couldn't read the data. Then I tried to break up the mirror and of course, I broke off the wrong drive, so I couldn't access the pool anymore (bc it only was the broken drive) and the good drive didn't seem to contain the data anymore? Also tried to read the data from CLI, that also didn't work..
TrueNAS seemed to use full resources to read the data when I imported the pool and after that I couldn't do anything, I've once waited days but nothing changed.

I've looked into programs that read blocks of data, but couldn't find one that did that. I even thought that maybe both drives failed at the same time..
As you can ready above, not really experienced but also a combination with stress, fatigue and just burned out from everything that I just started doing stuff hoping it would be all over soon.. The idea of losing my most important data while I actually thought I fixed the infrastructure this time is just..

The question here really is, how f am I. Is there a way to read blocks or raw data from this file system? Also, no snapshots anymore.. If I were to bring my drive to a data recovery company, could they guarantee to get my data back and would that cost hundreds?

Thanks in advance, I know this post is a combination of vagueness and stupid decisions, but I really want to use my server again and just let this part behind me.

 

[JoeAtWork]

you could use ddrescue to copy the data from on of your drive to a new drive that you know has 0 issues. Sounds like both of your drives have checksum issues

 

[artlessknave]

it's very hard to recommend anything without hardware specs. most of the forum members who CAN help with this won't if you dont, at the very least, follow the forum rules.
if you were using SMR drives, for example, they perform like utter garbage with zfs.
overall, ZFS has very little recovery. ZFS is an enterprise solution, and assumes you have backups to restore from. snapshots are not backups.
also, scale is not yet a great choice for data you care about that has no backups.

data recovery company

more like hundreds to thousands, though mirrors are fairly simple and might get a not super high price.

 

[Keyakinan]

you could use ddrescue to copy the data from on of your drive to a new drive that you know has 0 issues. Sounds like both of your drives have checksum issues

I did use ddrescue, recovered 100% with no errors.. allright.
Mounted it and it gives exactly the same error.. so now I'm 100% sure its a config thing or something.
I've let this run previously for a long long time and it will not go beyond 0% and takes up all my cpu power that I give it and also keeps writing to disk?

For people googling this: alert.process_alerts pool.import_pool

 

[JoeAtWork]

is this the same version of TrueNAS so the ZFS pool version is identical?

are you installing on a new system and importing just one of the mirrors?

If DDrescue found no errors you may have a volume corruption issue. I have seen hard drives not dump the cache to disk when power cycled, this would be bad for you if this is the case.

 

[Keyakinan]

is this the same version of TrueNAS so the ZFS pool version is identical?

are you installing on a new system and importing just one of the mirrors?

If DDrescue found no errors you may have a volume corruption issue. I have seen hard drives not dump the cache to disk when power cycled, this would be bad for you if this is the case.

No it is not the same version.. Didn't know this was a thing. I don't know what version it was, but I tried it then and it also didn't work.
I am using a new system (proxmox with truenas and drive pass through) and yes I'm importing just one of the mirrors.

Volume corruption issue doesn't sound great.. I'm gonna google it but is there anything I can do right now?

 

[JoeAtWork]

Try bare metal first, passthru can be VERY tricky and even if you think it is working right you might have an issue. Did either disk in your mirror with good data show a checksum for writes?

 

[Keyakinan]

Try bare metal first, passthru can be VERY tricky and even if you think it is working right you might have an issue. Did either disk in your mirror with good data show a checksum for writes?

until today I always did baremetal but for ddrescue I used proxmox with ubuntu so i could easily passthrough. I'm not sure what checksum for writes mean? Right now the fully new disk with a 100% disk-to-disk copy of the working drive of the mirror is imported in truenas so it's def not a hardware related issue.
I'm pretty sure it's a pool thing.. Right now I just want to get all the data off the drive, backup and wipe both disks.

 

[JoeAtWork]

zpool status -v $poolname

root@freenasmini[~]# zpool status -v test2 pool: test2 state: ONLINE scan: scrub repaired 0B in 06:24:06 with 0 errors on Mon Feb 21 06:24:10 2022 config: NAME STATE READ WRITE CKSUM test2 ONLINE 0 0 0 mirror-0 ONLINE 0 0 0 574227e0-85d2-4330-9874-ad0310885245 ONLINE 0 0 0 a36b66cb-d307-4b4a-8b07-b047f1dc25f8 ONLINE 0 0 0 79c87489-5552-431f-ab93-65bd5f9f026f ONLINE 0 0 0 errors: No known data errors root@freenasmini[~]#