Grub minimal-bash - Need help to fix broken boot

[elbzone]

Hello,

hopefully, I've tried hard enough to search for any information solving my problem before writhing this thread.

Without my interaction, my TrueNas-Scale installation (22.12 Bluefin Release) did a shutdown and is not able to boot since than.

Following a screenshot from the boot state. Grub is there but does not know how to continue.

From the boot entries itself, it is still there. I am able to select the nvme drive with the two expected (two expected, right?) boot entries

My problem is, that I did not find any documentation how to fix this grub installation. There are a lot of default grub error fixing how to's like [this](https://christitus.com/grub-rescue/) one. But as you see, by hd3 has a efi labeled partition and that's it.

If you could guide me to the knowledge base how to fix this, I am more than pleased.
If no official documentation for this exists, I am willing to write it and push it to the right repository. All I need is a collection of command plus a little bit of description for it.

I even more pleased to save time if the only answer would be "reinstall and setup the whole system again". This is not my ideal case but if trying to rescue the current installation is taking more time than to install and setup again, I have to eat that cookie.

Small details for the system.

An ASRock board with am AMD Ryzen 8 core.
2 4 TB Seagate Sata HDD's in Raid Mirror
1 256 GB Intel ssd as cache device
1 256 GB nvme ssd as boot device
64 GB ECC Memory

All was working great so far.

My problem is, that neither `set root=...` nor `set prefix ..` is resulting in a working `insmod normal` or in a boot. I am stucked at this point and again, can not find anything in the official [documentation](https://www.truenas.com/docs/scale/gettingstarted/)

Best regards and thanks in advance,
elbzone

 

[Arwen]

Why don't you simply re-install and restore your configuration?

That would take perhaps 30 minutes.


Their are problems if you used encryption and don't have the key saved.
Or, if you did not save your configuration.

 

[elbzone]

My configuration was saved last 2023-02-23.

It feels strange that the default answer would be so windows like (no offense, really).

I was hoping and searching for a "this are the default steps, be agile and adapt where you need to"-how to.
Loosing part of grub should not be that unusual. As I am a kid using zfs on an arch linux as daily driver since early 201x, I can tell you stories about broken boot loaders ;-).

 

[elbzone]

So I've done a backup via clonezilla and did a new install followed by an import of my latest configuration `truenas-TrueNAS-SCALE-22.12.0-20230223184913.db`.
I can log in from the local console as root

Now I have multiple issues:

* My data pool is not mounted
* I've set datetime manually but neither https is working nor I am able to log in via webgui with my known users
* Since my data pool is not mounted, I am not able to log in with my "admin" user (his home is located on `/mnt/data-pool/truenas/home/<username>`

Again, I tried my best reading the documentation but could not find anything useful for me :-(.

 

[Arwen]

Yes, I understand that installing again seems strange. But, this is an appliance, designed to have it's entire configuration savable. While in MS-Windows re-installing may require hours of work to configure as it was before. (Even if you could remember all the changes you made...)

You should be able to login via the console. Then escape into a SHELL. From their, run these 2 commands:
zpool status zpool import
Their are sometimes reasons why a pool won't auto-import.

 

[elbzone]

@Arwen,

thank you very much. From that point of few, a fresh install makes totally sense.

One last question.

While fixing my installation, I've taken some notes. I would like to beautify them and share them. Can you please point me to the fitting repository where I can create a push request?

My raw and not beautiful are:

```
* backup disk by using clonezilla
* install new
* system -> general -> configuration -> import
* reboot
* console
* 4
* choose the not root user to disable 2fa
* reset password
* login to web gui
* dataset -> unlock dataset
* if unlock is not doable
* login as root on the machine
* check if there is a path/directory below the expected mount point
* remove that
* example
* given is that you have to pool `test` which is mounted in `/mnt/test`
* when logged in, you find a directory path called `/mn/test/truenas/home/myuser`
* execute ` rm -fr /mn/test/truenas` if you've checked that the path is empty
* reboot
```

Best regards and thanks so far and in advance

 

[Arwen]

I don't know of any place for your "pull request".

The official TrueNAS documentation, (and probably available for pull requests), likely has all the recovery steps in it.

Further, making a "backup disk by using clonezilla" is likely not relevant to TrueNAS. Yes, it can be useful to keep the original boot device in case you have an older saved configuration. That is because on occasion, it is possible to extract the latest configuration off a boot device. Even if it won't boot. Depends on what was lost / damaged.