rockybulwinkle
Dabbler
- Joined
- Aug 2, 2021
- Messages
- 25
First let me explain my setup:
I have two installations of TrueNAS. One is local (named Titan and mirrored pool "data0"), and another is remote (named Atlas and mirrored pool "data0-backup"). Of note, Atlas only has 4GB of RAM. I am well aware that this is not supported and is the next thing I'm going to upgrade, but I really don't think this is the cause of my issue today.
Titan normally replicates to Atlas.
Last week, Atlas had a boot drive failure. The boot-pool wasn't in a mirror and I had no backups. The drive still worked "well enough" to boot and create a backup of the configuration, so I made that, and reinstalled TrueNAS. In the process I upgraded Atlas from Core 12 to 13.
Now, where the trouble really starts:
This weekend, I made an initial backup of a family member's data to Atlas (I'm building them a NAS for christmas and wanted to have that initial backup pre-loaded).
When I got home, I tried to create a replica using the Web UI from Atlas to Titan. This was done through Titan's UI as a pull from Atlas. I think I misconfigured the replication's destination as my root dataset initially (I think this is when my data started getting deleted). I corrected it, and tried starting the replication again, which failed.
At that point I decided to make the replica using the command line, unaware that I had deleted a bunch of datasets on Titan. I *think* around 6GB was written to Titan before I noticed the missing datasets.
As soon as I noticed this, I killed the replication process from Atlas to Titan, and tried to export Titan's data pool. It wouldn't let me, because apparently I'd set the system dataset to my data pool instead of boot-pool... I moved the system dataset to boot-pool, tried exporting again, rebooted, finally exported.
I tried importing against the last transaction before the destruction, but it says "cannot import 'data0': one or more devices is currently unavailable".
I can still import data0 on Titan, but several datasets are gone, and all common snapshots for replication to Atlas are gone.
I have put a hold on all snapshots on Atlas. I have not lost any data yet as far as I'm aware, just damaged my local NAS.
Is there any way I can restore my datasets? I really don't want to have to restore from the remote backup because it has a super slow upload speed (home internet). Of course I will if I have to (either by uploading or by sneakernet).
I do have a single drive locally that can hold all of what's left of data0, though I'm going to hold off on trying anything more until hopefully someone experienced can give me advice.
I have two installations of TrueNAS. One is local (named Titan and mirrored pool "data0"), and another is remote (named Atlas and mirrored pool "data0-backup"). Of note, Atlas only has 4GB of RAM. I am well aware that this is not supported and is the next thing I'm going to upgrade, but I really don't think this is the cause of my issue today.
Titan normally replicates to Atlas.
Last week, Atlas had a boot drive failure. The boot-pool wasn't in a mirror and I had no backups. The drive still worked "well enough" to boot and create a backup of the configuration, so I made that, and reinstalled TrueNAS. In the process I upgraded Atlas from Core 12 to 13.
Now, where the trouble really starts:
This weekend, I made an initial backup of a family member's data to Atlas (I'm building them a NAS for christmas and wanted to have that initial backup pre-loaded).
When I got home, I tried to create a replica using the Web UI from Atlas to Titan. This was done through Titan's UI as a pull from Atlas. I think I misconfigured the replication's destination as my root dataset initially (I think this is when my data started getting deleted). I corrected it, and tried starting the replication again, which failed.
At that point I decided to make the replica using the command line, unaware that I had deleted a bunch of datasets on Titan. I *think* around 6GB was written to Titan before I noticed the missing datasets.
As soon as I noticed this, I killed the replication process from Atlas to Titan, and tried to export Titan's data pool. It wouldn't let me, because apparently I'd set the system dataset to my data pool instead of boot-pool... I moved the system dataset to boot-pool, tried exporting again, rebooted, finally exported.
I tried importing against the last transaction before the destruction, but it says "cannot import 'data0': one or more devices is currently unavailable".
I can still import data0 on Titan, but several datasets are gone, and all common snapshots for replication to Atlas are gone.
I have put a hold on all snapshots on Atlas. I have not lost any data yet as far as I'm aware, just damaged my local NAS.
Is there any way I can restore my datasets? I really don't want to have to restore from the remote backup because it has a super slow upload speed (home internet). Of course I will if I have to (either by uploading or by sneakernet).
I do have a single drive locally that can hold all of what's left of data0, though I'm going to hold off on trying anything more until hopefully someone experienced can give me advice.
Last edited: