[9.x][BVT] Launching of an instance fails on execution of qemu-img info
Affects | Status | Importance | Assigned to | Milestone | ||
---|---|---|---|---|---|---|
Mirantis OpenStack | Status tracked in 10.0.x | |||||
10.0.x |
Fix Committed
|
Medium
|
MOS Nova |
Bug Description
A few recent builds of 9.0-community.
https:/
https:/
failed on launching of an instance with the following error:
2016-11-19 04:34:55.805 23824 DEBUG oslo_concurrenc
2016-11-19 04:34:55.806 23824 DEBUG oslo_concurrenc
2016-11-19 04:34:56.866 23824 DEBUG nova.compute.utils [req-b823b3fc-
2016-11-19 04:34:56.867 23824 ERROR nova.compute.
Note, that it took `qemu-img info` 5.177s to return -9. We intentionally run this command with resource constraints set (see https:/
/usr/bin/python -m oslo_concurrenc
(max 1G of memory and 2 seconds of CPU time)
to prevent possible DoS attacks with malicious images, but in this case the image is correct - VMs are booted successfully in other OSTF test cases.
My current understanding is that sometimes a CI node a test environment is run on can be "slow" (disk? cpu?) and cause such errors.
Changed in mos: | |
status: | New → Confirmed |
importance: | Undecided → Medium |
assignee: | nobody → MOS Nova (mos-nova) |
milestone: | none → 9.2 |
tags: | added: area-nova |
tags: | added: on-verification |
One more recent failure.