Adding new computes makes cluster unusable
Bug #1260490 reported by
Aleksandr Shaposhnikov
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Fuel for OpenStack |
Invalid
|
Critical
|
Matthew Mosesohn |
Bug Description
Build 124.
Create cluster with 1 controller and 1 compute.
Deploy it. It should work fine.
Go and add additional compute node. Deploy. Everything looks like fine but compute nodes disappears from openstack's dashboard hypervisors section and didn't registered in nova services. nova-manage service list didn't show them.
Changed in fuel: | |
importance: | Undecided → Critical |
milestone: | none → 4.0 |
Changed in fuel: | |
assignee: | nobody → Matthew Mosesohn (raytrac3r) |
status: | New → In Progress |
To post a comment you must log in.
I did this with CentOS and RHOS and both work amazingly. I can see 2 computes in nova-manage service list and see both computes in Dashboard and can launch instances on both computes. Did you do this in Ubuntu? Can you post diagnostic snapshot?