Hi Krystian,

Progress update on the bisection tests - all five hosts have now been
running for over a week without a single panic:

- srv04: mainline v6.8.12 (neither commit) + `processor.max_cstate=1 
idle=nomwait` - **8 days 14h uptime** (previously panicked at ~3 days on this 
same kernel without those params)
- srv08: mainline v6.13.0 (neither commit) - 7d 15h uptime
- srv06: mainline v6.14.0 (commit 1 only) - 6d 15h uptime
- srv07: mainline v6.15.6 (commit 1 only) - 6d 14h uptime
- srv05: mainline v6.15.7 (both commits) - 6d 14h uptime

All hosts continue running their full production VM load (7-16 guest VMs
each) with no issues.

Given how long and variable the failure window has turned out to be in
this whole investigation (we've seen panics anywhere from a few hours to
25 days), I don't want to draw firm conclusions yet just from a week of
uptime on any of these. But srv04 with Viktor's boot parameters is now
well over twice as long as its previous crash time on the identical
kernel, which is at least a positive signal for that workaround.

I'll keep all five running and will let you know if/when any of them
panic, or once we're comfortable calling this inconclusive after a
longer stretch.

-- 
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

Reply via email to