yahoo-eng-team team mailing list archive

[Artom Lifshitz <1986764@xxxxxxxxxxxxxxxxxx>]

Hello, thanks for the bug report!

IIRC the host oom killer runs per NUMA node, so in this case this sounds
like expected behaviour: node0 is out of memory (32 + 64 is bigger than
64).

** Changed in: nova
       Status: New => Invalid

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to OpenStack Compute (nova).
https://bugs.launchpad.net/bugs/1986764

Title:
  when there are vms with numa config and vms without, OOM may be occurs

Status in OpenStack Compute (nova):
  Invalid

Bug description:
  1.my host has 2 numa nodes, 64G memrory each node
  2.create vm1 without numa config, 64G memory
  3.create vm2 with numa config(1 numa node 32G memory),finally the vm's memory is allocated on numa node0 of host
  4.run memory test process in vm1，and let it's memory full(from numactl -H of host, i found vm1 used 40G memory on numa node0 of host)
  5.run memory test process in vm1，try to let it's memory full
  6.the vm1's qemu process is crashed by OOM

To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/1986764/+subscriptions