SOLVED Cannot boot when disk added

[Biiscuit971]

Hello everyone,

I have been using TrueNas Core edition Community with a professional PC that serves as a server (optiplex 3040) connected to a storage bay via a PCIe card that has two SFF outputs. It is the QNAP TL-D800S bay.

Since I added disks to this array after a reboot, I have Truenas that freezes on a disk listing screen and does nothing.

If I disconnect the bay from the server, it boots normally. To spare you the details, I noticed that the numbering of the disks was changed when I plug in the bay.
The mirrored boot disks (ada0 and ada1) are now in ada7 and ada8. I thought that this was probably the cause.

I've looked really hard and I haven't found a way to freeze the names of the boot disks or at least make them not based on the ada numbering. I'm not a FreeBSD user so if anyone has a solution, it would be very kind of you.

Meanwhile my solution is as follows : I wait for the Truenas message to appear and I plug the bay directly and everything works fine.

But is not a solution

Thanks to you

 

[HoneyBadger]

QNAP TL-D800S

While this does appear to be a standard SATA JBOD unit, I have my doubts about the suitability of the QNAP PCIe card. From what I can see, it appears to be driven by a pair of ASMedia ASM1164 chipset behind a PCIe switch.

Is replacing this card with an LSI HBA such as a 9200-8e (or equivalent flashed OEM card, like the Dell H200e) an option?

 

[Biiscuit971]

I want to keep this PCIe card. The replacement is not envisaged for the moment.
Do you think the problem could come from there?

 

[HoneyBadger]

I want to keep this PCIe card. The replacement is not envisaged for the moment.
Do you think the problem could come from there?

It's possible. Assuming I'm correct in my assessment of the card from the QNAP website [1] it's going to present the drives as SATA devices, so if there's a BIOS on the QNAP card it might be trying in vain to boot from those disks, or somehow otherwise confusing the Optiplex.

Can you compare the Optiplex BIOS with respect to "boot order" and "SATA devices" when booting to BIOS with the bay attached vs disconnected? Checking for and disabling the "PCI Option ROM" for the add-in card (or "cards" if it shows as two devices) might also help here.

[1] https://www.qnap.com/en-me/product/tl-d800s

 

[Biiscuit971]

you can see my display when the bay is plugged in : (Samsung ... and CT250MX500 ... 2 SSDs that contain the mirrored system.)

In the boot sequence options, I have chosen the SSDs first but not working.

I don't know if it's this option but I can't disable it because I have TrueNAS installed in Legacy mode.

if I try to boot on the Truenas with USB bootable in legacy mode it also blocks but if I boot in uefi mode it works.

Reinstalling TrueNAS in UEFI mode solve this problem ?

 

[HoneyBadger]

Reinstalling TrueNAS in UEFI mode solve this problem ?

It might, if the ASM1164 chip on the QNAP card can't do UEFI boot then it might just default to only the drives it can see (on the onboard Intel based SATA ports) but I can't guarantee that. You could also use a USB-to-SATA adapter and put the USB boot device first on the list to see if that will take priority.

A bigger problem I see though from your BIOS images is your hardware - three of your four 4TB drives are WD40EFAX which are known to use shingled magnetic recording or "SMR" - these drives specifically have firmware compatibility issues with ZFS that can cause them to fail during heavy writes, including during resilvering ("rebuild") operations.

WD Red SMR Drive Compatibility with ZFS

Thanks to the FreeNAS community, we have uncovered a potential ZFS compatibility issue with some capacities of newer WD Red drives that use SMR (Shingled Magnetic Recording) technology. Update to this post...

www.truenas.com

You should try to plan for their replacement.

 

[Biiscuit971]

I'll try it all this weekend.

And thank you for the information about the disks I did not know.

 

[Biiscuit971]

If I limit the write speed on the zfs pool, can this reduce the risk?

While waiting for the replacement

I have just seen the recommendations if we have SMR disks I will apply them now

 

[Biiscuit971]

Hello,

I reinstalled TrueNAS in UEFI mode by choosing the upgrade option.
I deactivated the legacy rom option in the bios and it works.

For the hard disks I will see if it is possible to change them.

Thanks for the help
I