Appliance stuck with POST error B4 after upgrading

[dj0Nz]

Hi Mates,

about 50% of our 6000 series appliances (verified with abount 20 cluster up to now) are not coming back in time after upgrading them. They get stuck with POST code "B4" and (sometimes) continue booting after 2 to 4 hours. You can shorten the wait time by unplugging the power cords, waiting a moment, and then plugging them back in.

Before opening a TAC case:

Has anyone here ever had a problem like this?

Are there any known issues with firmware or LOM firmware on these appliance types (LOM is not used currently)?

Thanks in advance!

Regards
Michael

[Shyyyy]

Could you share maybe a little more information? screenshots?

[Chris_Atkinson]

Is the LOM firmware up to date, can you test a boot post updating?

What is the upgrade process / path being undertaken (Source / Target versions etc)?

Independently of the upgrade do they boot without delay, any USB devices connected?

[dj0Nz]

The systems are booting normally. Only updates are triggering these "boot delays". And it seems like it doesn't matter, which upgrade has been performed. It started with one of the later (not necessaryly the latest) R81.20 JHF und is still there with R82 major upgrade and R82 JHF installations.

USB devices are not connected. LOM is not in use on these appliances, but I checked the version using ipmitool and there is a very old Java-Based LOM version installed, which I'm suspecting to be the culprit (see https://www.truenas.com/community/threads/bios-hangs-on-b4.34856/ which is describing similar symptoms).

Problem is: These devices are located throughout the country and for regulatory reasons, they're not easily accessible. To make things worse, the error cannot be reliably reproduced.

So I just wanted to know if any other one than me had this problem before and maybe solved it or TAC solved it and a reason is known...

[Chris_Atkinson]

There are some cases I've seen where RMA'ing the LOM card has resolved similar issues. 

Do you have the Java or HTML5 version?

[Lesley]

Is the BIOS up to date, and same on all gateways? -> https://support.checkpoint.com/results/sk/sk120915

Make sure LOM is up to date: -> https://support.checkpoint.com/results/sk/sk88064

Even tho LOM is not used up date them because they can cause delay during boot

Check what the difference is betweens table 6000 and unstable, maybe one has HTML lom the other one Java? Or BIOS version? More interfaces or different interface cards? Maybe LOM still has ISO loaded? boot order also OK? 

[dj0Nz]

Thanks Lesley. Will check BIOS versions tomorrow. I suspected LOM because B4 may be caused by PCI/USB init problems, but we also have affected appliances without LOM. And: There was no problem before R81.20. I bet there was something like a firmware upgrade or modified kernel options that came with R81.20. TAC case is open. We'll see.