Truenas Scale - Boot Pool Failure, EDID 0 Errors

Computer Hardware & Server Infrastructure Builds
1

Okay techies got a question for those that have run into this exact problem.

Running 2 instances of Truenas:

1-Truenas Core running bare metal on a Supermicro with an Intel Xeon and On board Intel Graphics, ECC etc. It’s a server-based board.

1-Truenas Scale running inside Proxmox on another Supermicro board with hardware pass through and no issues, other than I need to manually start it versus automatically starting. Not a big deal as this is mostly a backup/r-sync machine.

Now when I tried to install Truenas Scale Electric EEL, Dragonfish, on the bare metal, I continually got “boot pool import” error then drops to (initramfs).

I double checked my SSDs for overall health, read, write and for errors. They are clean, at least by the software they show clean and healthy. I even deleted and wiped the partitions on both of them after doing the drive integrity check.

At one point I was able to get the installation to boot after some command line examples I found, but then I started getting the EDID 0 error based on the iGPU and i915 driver issue. I made appropriate changes, yet that issue continued to prop up on the console.

I disabled just about everything for the Intel GPU leaving only the ASPEED chip for display.

This is part of the reason I stuck with Core, because, well it just works out of the box. This is the second attempt at trying to run Scale on a bare metal, and it hasn’t worked out to well for me anyways. Now virtualized Truenas Scale works, which boggles me.

This was brought up in the iX Systems bug tracker and was supposedly addressed, but there are other examples of people continuing to run into this even after they acknowledged it to this day.

Any ideas and experiences on installing Truenas Scale without having to address these issues?

From BIOS settings to UEFI parameters.

Command line input

Etc.

I have nothing but respect for iX and the software, yet CORE works, Scale Continues to be buggy. I would like to have a bare metal running Truenas Scale since they are deprecating Core, but it’s the same issues after Scale first came out.

Thanks

2

Why aren’t you trying to install their latest 25.04 Fangtooth?

3

Tom,

Thank you for replying.

I did try that, though I can not remember what order I tried them. I believe I started with Fangtooth, then I started to step down to see which one would work based on what I read from other individuals in the Truenas forums and other forums based on the Kernel issues and boot pool error leading into (initramfs).

The EDID 0, no matter what I tried still didn’t seem to fix the problem, which resulted in continued CPU spikes and eventually machine lock up. The only way to clear the system was to use the IPMI and remote reset, or turn off then back on at the power supply. This occurred when Scale first came, which is why I chose to stick with Core.

The only thing I have not tried was doing an in place upgrade from Core to Scale.

At the moment I have not made any major changes to Core yet outside of setting up an r-sync peer to peer on separate nics from the primary nics on the MB for migrating my data back to Core.

I have always used Supermicro boards as they tend to be to the least buggy, error free of the server platforms, and are not proprietary. Any other Software/OS platform works fine with out issue. Even putting a version of Server 2025 works fine outside of having to upgrade the iGPU drivers on the CPU.

So I don’t know. I would have to think iX is aware of this still occurring, and its just a matter of coding it out.

The only thing I have not done hardware wise was replace the SSD’s with a new matched pair. I have used these since the days of Freenas and they continue to work fine. Again even putting core back on them, the system just works.

I am not knocking iX as they have put out a great NAS OS compared to others I have tested and used. I am just frustrated that this is a big enough issue that others randomly run into this whether one or the other, or even both.

Again, Thanks for the reply. Maybe I am missing something simple, which is why I walked away for a while and see if anyone else has had the problems and a resolution.

Robert

4

I have several systems running Supermicro boards and some of IX Systems box ship with them. This is not an issues I have run into and is since it’s probably uncommon it’s why they have not fixed it. TrueNAS is based on Debian which works well on a variety of hardware.

5

Tom,

I agree with everything you mentioned.

I have not ever had a problem running Debian on anything myself.

So my THEORY, putting aside Truenas as the culprit:

The SSD’s being older, may just be on there way out regardless what the software tests show. Could be boot loader is having conflicts/or issues reading the controller chip/firmware etc, or the DRAM cache chips are failing.

My plan is to try and install Scale again on a different SSD, maybe using a spare M.2 since the Supermicro board supports it, to figure out the Boot Pool Error. I would prefer to use 2.5 SSD, and save the M.2 for something like a cache drive for METADATA as the server has 64gigs, again its just a test.

On the EDID 0. I’ve seen conflicting comments on it being a monitor issue to a GPU/iGPU issue, to being a Kernel issue on different distros of linux. The problem shows up even on the Unraid forums, though supposedly a newer Kernel fixed this. I am going to try and install Scale again using a different monitor, and/or, remote install using the IPMI onboard KVM, then if that fails, maybe replace the Xeon w/iGPU with a non iGPU.

It would be nice if Truenas did allow you to isolate hardware from the GUI so you can try different troubleshooting methods without command line, but, I get it.

Thanks

Robert

6

Update:

So I found that the SSDs were in fact bad and showing inconsistent information on various programs and health checks. After replacing them with newer ones and doing a mirror install, I am successful at having a bare metal Truenas Scale Install with any errors at all.

Delving further into this, it seems when the controller chip and software don’t align with smart reporting errors it can have consequences and installation problems.

Now if I can figure out how to do a peer to peer rsync to make an exact copy over 10gig nics I’ll be able to finish this data migration.

Each NAS has 2 onboard 1gig nics for internet/network connection, and 1 10gig card for peer to peer data transfer to by pass the switch and router, each with its own static I.P.

I did read and try setting up rsync according to the Truenas guidelines but must be missing something cause it didn’t take.