Hi Krystian, Viktor, Status update after roughly 2.5 weeks of the bisection round (as of 2026-09-28), all hosts still without a panic:
- srv04: v6.8.12 (neither commit) + `processor.max_cstate=1 idle=nomwait` - 18d 19h - srv08: v6.13.0 (neither commit) - 17d 20h - srv07: v6.15.6 (commit c861cac950fb only) - 16d 20h - srv05: v6.15.7 (both commits) - 16d 19h - srv06: HWE 7.0.0-31 (moved there after its two panics on v6.14.0) - 6d 2h Two observations: 1. srv08 on v6.13.0 (no fixes, no boot params) has been running as long as the hosts with the fixes, so uptime alone does not separate the candidate patches. I agree with Viktor that this is not proof against a low-frequency race, but it also means these runs cannot confirm the patches either. 2. The only direct reproduction on a kernel without the fixes was srv04 on v6.8.12 (panicked after about 3 days, with the migrate_swap/multi_cpu_stop trace). I have not done a second run of v6.8.12 without the boot params, so I cannot tell whether that 3 days was typical or bad luck. Same for srv04 with the params: 18+ days is encouraging, but a single run. If it would help, I can switch srv04 back to plain v6.8.12 without the boot params for a second baseline run. That is probably the cleanest way to tell whether the params or the timing explain the difference. Let me know which you prefer. -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/2163642 Title: Kernel panic "Attempted to kill the idle task" on AMD Opteron multi- node NUMA under KVM (6.8.0-137) To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2163642/+subscriptions -- ubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
