Validation should catch unavailable resources
Bug #1388047 reported by
Steven Hardy
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
OpenStack Heat |
Fix Released
|
High
|
Kanagaraj Manickam |
Bug Description
If you try to launch a stack containing, e.g a OS::Neutron resource on an openstack which doesn't have Neutron, we fail to catch that during validation and start the stack create, then fail later.
It would be better if we could validate availability of the resources (by checking the client can be instantiated, this the endpoint is found?) such that we fail fast when a service is not available.
Possibly, we could add a test to the resource superclass which tries to instantiate the resource client?
Changed in heat: | |
assignee: | nobody → Tetiana Lashchova (tlashchova) |
Changed in heat: | |
importance: | Undecided → High |
milestone: | none → liberty-2 |
status: | In Progress → Triaged |
Changed in heat: | |
assignee: | Tetiana Lashchova (tlashchova) → Kanagaraj Manickam (kanagaraj-manickam) |
status: | Triaged → In Progress |
Changed in heat: | |
status: | In Progress → Fix Committed |
Changed in heat: | |
status: | Fix Committed → Fix Released |
Changed in heat: | |
milestone: | liberty-2 → 5.0.0 |
To post a comment you must log in.
Once I had an idea of detecting if Neutron is available during heat-engine start and simply not to register Neutron resources then (so they do not even show up in heat resource- type-list) , but it turns out some deployments like TripleO do start heat before Neutron...
Also, especially for Nova-network/ Neutron case, we have some resources that are "transparent" and use several clients, so there is no single "resource client".