We got a few new machines: x3850 x6.
All could pxe boot fine, except one machine, that gives the following kernel panic, looks like an exciting issue:
We cannot even scroll up after the kernel panic occurs, after 30-40 seconds. It hungs so bad, that I cannot even type anything.
Anybody have any clue, what could the problem possibly be? If it is a HW bug, then what to replace? CPUs? Motherboard?
- the BIOS settings are the exact same vs. working ones
- the firmware/bios versions are the exact same vs. working ones
- tried cold boot, the same kernel panic
- tried to boot with kernel parameters: "acpi=off" - it just did the same kernel panic at ~18 sec, not the usual panic at 30-40 sec.
- tried: "noapic nomodeset xforcevesa" - panic after 30-40 sec.
- tried: "acpi=off noapic nomodeset xforcevesa" - panic after 30-40 sec.
- tried: "isolcpus=0" boot param, same kernel panic, after 30-40 sec.
- tried to boot a slacko-5.6-PAE.iso - it booted normally! 3.10.5 SMP PAE. But we have to use SLES. The PAE kernel only sees ~65 GByte RAM, if that is a useful info.
- tried: https://www.memtest86.com/downloads/memtest86-iso.zip to run a simple memtest, but after 59 seconds of run without memory error, the machine freezed. -> UPDATE: The Memtest86+ from: http://www.memtest.org/#downiso doesn't freezes.
Once I seen: "Kernel panic - not syncing: Watchdog detected hard LOCKUP on cpu 18" - there are 4 CPUs in the machine, each has 18 cores, so don't know which one is this..
UPDATE: with the "maxcpus=0" kernel boot parameter, it finally booted, but still investigating, because still said: "A start job is running for Activation of LVM2 logical volumes (Xmin xs / no limit)" - but maybe CPU HW issue?