flexible use of cloud backend
As an openQA infrastructure manager I want to use openstack for worker nodes to have a more scalable openQA environment
- AC1: Cloud nodes are created on-the-fly when openQA jobs using this backend are used
- AC2: If cloud capacity is exceeded openQA jobs will stay scheduled until the next cloud node can be created and used
- create cloud node using openstack API when test run using openstack backend is started
- check cloud capacity and catch node creation errors properly (see AC2)
- bootstrap test on this node (see #12158)
- destroy cloud node after finish of test run
- optional: Extend openQA API/UI to configure parameters, e.g. limit of nodes to create
Idea is to implement a flexible cloud backend in os-autoinst using openstack API. The use of proper cloud could remedy problems when running into capacity limits, etc. This would also need better feedback in the UI in case nodes can not be created or capacity is depleted.
#6 Updated by okurz about 1 year ago
- Subject changed from [tools]flexible use of cloud backend to flexible use of cloud backend
- Status changed from New to Workable
- Difficulty set to hard
By now we haven't found further need for a specific cloud backend. Probably nowadays there might also be an alternative with k8s anyway. Of course there is no problem in running openQA on cloud instances per se. I still consider the original idea nice but not on the backlog of the QA tools team for now.