Problem Installing DSM7.2.2 on HP N40L Microserver

[Dragsdal]

Hi guys,

 

I think I've might have destroyed my installation trying to update, and need some advise.

 

I had DSM 7.2.1 69057-Update 5 running on a HPE ProLiant MicroServer N40L with TCRP 0.10.0 as loader.

I previously had succes manually updating through all the updates from DSM control panel, and was comfortable trying that once more as I saw DSM 7.2.2-72806 was avaliable.

The update seemed to go fine, but i ran into the infamous infinite recovery loop, which I had experienced before.

At the previous time, all I had to do was update the loader, and it usually did the trick.

 

This time there wasn't any update to TCRP so I did some reading and saw that someone had success updating a N40L to DSM 7.2.2-72806 with the RR loader, so I downloaded and flashed it (24.10.3).

The upgrade seemed to go fine, but to my horror I wasn't presented with the migrate option, but the "new install process" instead (naming the system, admin user, etc.)

 

When I finally got into DSM it was a brand new setup. Everything was back to factory default, and none of my apps/programs was there.

In the drive settings I could see the hard drives, bur each of them is having "System partition failed" as status...!

I'm being presented with the option to do a data scrubbing to ensure data consistency but I dare not do it, as I fear of erasing every single hard drive.

 

Can anyone please help, as I'm at the ropes here... 

 

Sincerely,

Dragsdal

 

 

 

Edited November 2 by Dragsdal

[Peter Suh]

My TCRP-mshell, which is a more advanced version of pocopico's TCRP, supports HP-N36L/N40L/N54L.
I'm sorry that the configuration information was lost due to the use of rr.

 

If you have an environment that you want to try again from the backed up original, download the latest version of the img file in my profile,

and be sure to change the default FREIND mode to JOT mode when building the loader to build the desired version of the loader.

TCRP's Jot mode and RR's direct boot mode prevent kernel panic from the latest buildroot kernel due to the use of old AMD CPUs.
(Skip the buildroot kernel and go straight to DSM kernel loading.)

 

Please note that UEFI boot boards are not supported in Jot mode.
Be sure to change to LEGACY.

 

I just re-ran the migration test from 7.2.1 to 7.2.2 on my DS918+ in JOT mode.

 

It prompts me to install the new 7.2.2 version,
and during this process it starts by saying it has detected a new DS918+.
Then it asks me if I want to proceed with the migration.

 

 

Edited November 3 by Peter Suh

[Dragsdal]

Hi Peter,

 

Thank you so much for your reply.

 

Following your guide, I am able to update from DSM 7.2.1-69057 to DSM 7.2.2-72806 (This was possible with RR as well to be fair).

The problem is however, that it boots right into the DSM 7.2.2-72806 enviroment and does not recognize that I came from a running system on 7.2.1-69057.

But there is some good news: the harddrives no longer registers af "System partition failed" but are now all "Healthy".

Even more good news is that all my data seems to be avaliable from disk station.

 

The major issue now is, that the entire DMS OS is back to factory default, so all apps (Docker, Resillio Sync, Survelliance Station, etc.) are gone.

It's not the worst problem, as I am able to configure everything back again - it's just a time consuming task I'd rather wish I hadn't do.

 

If you or other know a way of recovering the OS configuration I'd appreciate it.

 

But a big thanks to you Peter for helping me out so far. I owe you a beer!

 

Sincerely,

Dragsdal

Edited November 3 by Dragsdal

[Peter Suh]

If you want to restore the package configuration, you should have a backup of that as well.

If you want to restore other settings, you should also have a backup of those.

The image above is in Danish, so it's hard to understand,

but does it refer to a previous situation where the system partition was problematic? Is that fixed now?

 

It will probably take a while to resolve.
Once the RAID rebuild is complete, your data will probably be usable again.
Please make sure there is absolutely no power outage during this rebuild period.

Edited November 3 by Peter Suh

[Dragsdal]

I'm afraid I haven't got a backup of the package configuration. 😕

 

Most of my apps was running as docker containers anyway, and the folder containing all the config-folders for my docker containers is saved, so I guess it's just a matter of creating all the containers again and linking the corresponding config-folder to each. 

 

The image in danish basically says that all disks are OK, but warns that I need to run an optimization (data scrubbing ) to ensure data consistency.

I kind of feel confident that it won't ruin anything as all the data on storage pool 1 is intact.

So yea, it appear to be fixed as far as recovering the data.

I just need to reestablish the DSM OS (packaged, docker, etc.).

 

Sincerely,

Dragsdal

[bearcat]

On 11/3/2024 at 4:43 AM, Peter Suh said:

change the default FREIND mode to JOT mode when building the loader

 

Just curious, as I have been given the task to "save" a friends N40L (DS3622xp+) based on pocopico, with the same "update" problem as @Dragsdal experienced.

 

 

How/where do you change this during the process?

On 11/3/2024 at 4:43 AM, Peter Suh said:

change to LEGACY.

 

The same question here, where and when do I change it?

 

Edit:

Started fresh with a blank USB, and it "just worked" after the first reboot

Edited November 8 by bearcat

[Peter Suh]

40 minutes ago, bearcat said:

 

Just curious, as I have been given the task to "save" a friends N40L (DS3622xp+) based on pocopico, with the same "update" problem as @Dragsdal experienced.

 

 

How/where do you change this during the process?

 

The same question here, where and when do I change it?

 

Edit:

Started fresh with a blank USB, and it "just worked" after the first reboot

 

 

There have been some recent improvements to support Jot.
Please use the latest version of mshell in my profile, 1.0.5.2.
Now, both UEFI and Legacy support Hybrid GRUB.

You no longer need to change it to Legacy.
Please use either UEFI or Legacy, whichever is more convenient.

 

If you are stuck in an infinite recovery loop, you must process PostUpdate if you have used a patch after Update 1,

which was released after the distribution of DSM.

Without this PostUpdate, the infinite recovery loop will repeat.

 

First, build a new Jot mode mshell 1.0.5.2 loader according to the version of DS3622xs+ you want,

and use the first PostUpdate in the GRUB menu of the first image I captured and uploaded.

You can escape the infinite recovery loop.

 

Originally, if a version after Update 1 is used, mshell automatically handles PostUpdate internally.
You can think of the first Grub menu as an emergency.
That's why one automatic reboot occurs, and it feels like "it just worked."

Edited November 8 by Peter Suh