@mluypaert

Thanks for the details.

Do you know whether the problem reproduces with any bowtie2 example that
I could run for myself?  I'm not familiar w/ it.

Apparently there's a workaround for it, if you're willing to test: to
disable NUMA balancing.

This _might_ impact performance on some memory-intensive workloads that use 
memory on multiple NUMA nodes.
If you previously used kernel versions 3.13.0-156 to 3.13.0-158 there should be 
no performance impact, as NUMA balancing was disabled in those versions to 
workaround a regression on 3.13.0-155. It was re-enabled by default on 
3.13.0-159.

You can disable NUMA balancing on a running system with this command:

$ sudo sysctl -w kernel.numa_balancing=0

And then make the change persistent across reboots with this command:

$ echo 'kernel.numa_balancing = 0' | sudo tee /etc/sysctl.d/99-numa-
balancing.conf

If you'd like to try that, please let us know how it goes in a few
months maybe? :-)

Thanks for your response anyway.
cheers,
Mauricio

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1813018

Title:
  Kernel Oops - unable to handle kernel paging request; RIP is at
  wait_migrate_huge_page+0x51/0x70

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1813018/+subscriptions

-- 
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to