[Upgrade] Re-deployment of simple 5.1.1 cluster failed after upgrading to 6.0
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Fuel for OpenStack |
Fix Released
|
Critical
|
Ihor Kalnytskyi | ||
6.0.x |
Fix Released
|
Critical
|
Ihor Kalnytskyi | ||
6.1.x |
Fix Released
|
Critical
|
Ihor Kalnytskyi |
Bug Description
{
"build_id": "2014-12-
"ostf_sha": "a9afb68710d809
"build_number": "50",
"auth_
"api": "1.0",
"nailgun_sha": "dc621c96e30f0d
"production": "docker",
"fuelmain_sha": "fb898768123204
"astute_sha": "16b252d93be6aa
"feature_
"mirantis"
],
"release": "6.0",
"release_
],
}
},
],
}
}
},
"fuellib_sha": "484f53f586c084
}
Steps:
1. Create next fuel 5.1.1-48 cluster - Ubuntu, simple, Neutron Vlan, ceph for volumes, ceilometer, sahara, 1 controller+mongo, 2 compute nodes, 2 ceph nodes
2. Deploy cluster
3. Upload 6.0-50 upgrade script to master
4. Upgrade fuel to 6.0
5. Add 1 compute node to cluster
6. Re-deploy
Actual - re-deployment failed, error in puppet.log of node-6
Wed Dec 17 14:54:12 +0000 2014 Puppet (err): Invalid parameter other_nets on node node-6.
Logs are attached
Changed in fuel: | |
status: | New → Confirmed |
importance: | High → Critical |
tags: | added: docs |
It looks like the puppet manifests for the upgraded env is correct 5.1.1, but the astute.yaml generated includes several extra parameters, including:
other_nets
default_gateway
It looks like we're using the wrong deployment serializer for deploying node-6.