Deployment of 5 controllers failed with errors in puppet log
Bug #1529874 reported by
Egor Kotko
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Fuel for OpenStack |
Invalid
|
High
|
Bogdan Dobrelya |
Bug Description
Deployment of 5 controllers failed with errors in puppet log
Scenario:
1. Create new environment
2. Choose Neutron, VLAN
3. Choose Ceph for volumes and images, ceph for ephemeral
4. Add 5 controller
5. Add 2 compute
6. Add 2 ceph nodes
7. Change management net default mask from /24 to /25
8. Change default disk partitioning for ceph nodes for vdc
9. Verify networks
10. Deploy changes
Actual result failed with errors in puppet log:
http://
Diagnostic snapshot:
http://
To post a comment you must log in.
Here are events flow http:// pastebin. com/fMZSvKLh
The dump definitions task was failed as the node-19 was running with dead rabbitmq node. The deploy was failed because of the network partitions ended up with few nodes dropped dead permanently and was never back in rabbit cluster. There is no commands output so I cannot say what was the state of the pacemaker cluster, likely the dead / split as well. So this bug is rather invalid