juju deployment of systems that does not meet maas base constraints
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
juju-core |
Triaged
|
Critical
|
Ian Booth |
Bug Description
When doing an OpenStack deployment, our maas base constraints of hw-staging was totally ignored. All the servers selected for the bootstrap node and the OpenStack deployment did not meet the constraints.
Some console output showing maas base tag:
=======
2015-01-14 22:55:14,199651: Base MAAS tag: hw-staging
2015-01-14 22:55:14,200009: converted slave label ci-oil-slave1 to maas user oil-slave-1
2015-01-14 22:55:15,111365: found 34 nodes with tag=hw-staging
2015-01-14 22:55:15,111848: restricting to 0 nodes owned by oil-slave-1
2015-01-14 22:55:15,112040: no nodes owned by oil-slave-1 in tag hw-staging
Exiting with: 0
+ rm -rf /var/lib/
+ mv run_config artifacts/
+ oil-ci generate run_config --out run_config --environment ci-oil-slave1
2015-01-14 22:55:15,548 [INFO] oil_ci.cli: Constructing parameters from jenkins job environment
2015-01-14 22:55:15,549 [INFO] oil_ci.cli: Generating run config with deploy parameters: {'block-storage': 'cinder-iscsi', 'networking': 'neutron-
2015-01-14 22:55:15,549 [INFO] oil_ci.cli: Generating run_config networking for env: ci-oil-slave1
2015-01-14 22:55:15,549 [INFO] oil_ci.cli: Allocating networks for env: ci-oil-slave1
2015-01-14 22:55:15,603 [INFO] oil_ci.cli: OIL Network Allocating 2 ranges of 32 IPs per environment
2015-01-14 22:55:15,618 [INFO] oil_ci.cli: Calculating deployment cost
2015-01-14 22:55:15,914 [INFO] oil_ci.
2015-01-14 22:55:15,915 [INFO] oil_ci.
2015-01-14 22:55:15,915 [INFO] oil_ci.
2015-01-14 22:55:15,915 [INFO] oil_ci.
2015-01-14 22:55:15,915 [INFO] oil_ci.
2015-01-14 22:55:15,915 [INFO] oil_ci.
2015-01-14 22:55:15,916 [INFO] oil_ci.
2015-01-14 22:55:15,916 [INFO] oil_ci.
2015-01-14 22:55:15,916 [INFO] oil_ci.
2015-01-14 22:55:15,916 [INFO] oil_ci.
2015-01-14 22:55:15,916 [INFO] oil_ci.
2015-01-14 22:55:15,916 [INFO] oil_ci.
2015-01-14 22:55:15,917 [INFO] oil_ci.
2015-01-14 22:55:15,922 [INFO] oil_ci.
+ cp -av run_config artifacts/
`run_config' -> `artifacts/
+ cp -av oil_nodes artifacts/oil_nodes
`oil_nodes' -> `artifacts/
+ oil-ci do deploy --environment ci-oil-slave1 --run_config artifacts/
2015-01-14 22:55:16,753 [INFO] oil_ci.
2015-01-14 22:55:17,011 [INFO] oil_ci.juju.client: Boostrapping new environment
Bootstrapping environment "ci-oil-slave1"
Starting new instance for initial state server
Launching instance
WARNING no architecture was specified, acquiring an arbitrary node
- /MAAS/api/
Installing Juju agent on bootstrap instance
Waiting for address
Attempting to connect to mokoi-vm-02.oil:22
Attempting to connect to mokoi-vm-02.oil:22
=======
For example (I'll attach entire yaml file as well):
=======
environment: ci-oil-slave1
machines:
'0':
agent-state: started
agent-version: 1.21-beta4
dns-name: mokoi-vm-02.oil
hardware: arch=amd64 cpu-cores=2 mem=12288M tags=virtual
instance-id: /MAAS/api/
series: precise
state-
'1':
agent-state: started
agent-version: 1.21-beta4
containers:
1/lxc/0:
dns-name: 10.245.44.246
hardware: arch=amd64
series: trusty
1/lxc/1:
dns-name: 10.245.44.248
hardware: arch=amd64
series: trusty
dns-name: zahhak.oil
hardware: arch=amd64 cpu-cores=4 mem=32768M tags=hardware-
instance-id: /MAAS/api/
series: trusty
=======
We are using juju 1.21-beta4 and maas version 1.7.1~rc1+
Note that this issue has not been observed with Juju 1.20.x and maas version 1.7.1~rc1+
juju-status.yaml file