QEMU >= 5.0.0 with -accel tcg uses a tb-size of 1GB causing OOM issues in CI
Bug #1949606 reported by
Lee Yarwood
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
OpenStack Compute (nova) |
Fix Released
|
Medium
|
Lee Yarwood |
Bug Description
This is a Nova tracker for a set of issues being seen in OpenStack CI jobs using QEMU >= 5.0.0 caused by the following change in defaults witin QEMU:
https:/
https:/
At present most of the impacted jobs are being given an increased amount of swap with lower Tempest concurrency settings to avoid the issue, for example for CentOS 8 stream:
https:/
https:/
Longer term a libvirt RFE has been raised to allow Nova to control the size of the cache:
description: | updated |
Changed in nova: | |
status: | Confirmed → In Progress |
To post a comment you must log in.
Putting to Confirmed given we already triaged it.