Some Tempest tests require booting multiple instances at once, but we only deploy 1 Ironic node
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Ironic |
Fix Released
|
Medium
|
Unassigned | ||
OpenStack Core Infrastructure |
Triaged
|
Medium
|
Unassigned | ||
tempest |
Won't Fix
|
Undecided
|
Unassigned |
Bug Description
Some compute API tests require booting multiple instances during a single test or single test case, ie:
tempest.
tempest.
tempest.
tempest.
The current Ironic devstack-gate deployment sets up Ironic with a single VM enrolled as a single Ironic node. Once this has been provisioned, there are no more resources available for additional test instances.
Some solutions:
-Tempest needs to have a configurable limit to the number of instances that can be spawned at once
- Ironic+Devstack deployment should set quota limits based on IRONIC_VM_COUNT
- Adjust IRONIC_VM_COUNT in the devstack-gate job to match needs of Tempest.. not sure how well the devstack slave nodes would handle deploying many Ironic instances. tempest.
Changed in ironic: | |
status: | New → Triaged |
importance: | Undecided → Medium |
tags: | added: ironic-tempest |
Changed in openstack-ci: | |
status: | New → Triaged |
importance: | Undecided → Medium |
Changed in ironic: | |
status: | Triaged → Fix Released |
This issue should be solved on the ironic side. Tempest assumes a modest level of resources which should be met when using it. The multi-server tests can be skipped if running in a limited server environment.