Attempting to run charm before unit provisioned, 1.26

[Stuart Bishop]

Here is status ouput showing that postgresql/17 running on machine 26 has failed. It likely failed because it doesn't have an IP address and the machine clearly listed in 'pending' state.

[Services]
NAME STATUS EXPOSED CHARM
client unknown false local:trusty/pgclient-7
postgresql error false local:trusty/postgresql-11

[Units]
ID WORKLOAD-STATE AGENT-STATE VERSION MACHINE PORTS PUBLIC-ADDRESS MESSAGE
postgresql/17 error idle 1.26-alpha3.1 26 5432/tcp hook failed: "db-relation-changed" for client:db

[Machines]
ID STATE DNS INS-ID SERIES AZ
0 started 10.0.3.145 juju-5f65e7eb-5d0b-4c6e-8fde-143ad97856fc-machine-0 trusty
26 pending juju-5f65e7eb-5d0b-4c6e-8fde-143ad97856fc-machine-26 trusty

[Stuart Bishop (stub)]

environment: lxd
machines:
  "0":
    agent-state: started
    agent-version: 1.26-alpha3.1
    dns-name: 10.0.3.145
    instance-id: juju-5f65e7eb-5d0b-4c6e-8fde-143ad97856fc-machine-0
    instance-state: Running
    series: trusty
    hardware: arch=amd64 cpu-cores=0 mem=0M
    state-server-member-status: has-vote
  "26":
    agent-state: pending
    agent-version: 1.26-alpha3.1
    instance-id: juju-5f65e7eb-5d0b-4c6e-8fde-143ad97856fc-machine-26
    instance-state: Running
    series: trusty
    hardware: arch=amd64 cpu-cores=0 mem=0M
services:
  client:
    charm: local:trusty/pgclient-7
    exposed: false
    life: dying
    service-status:
      current: unknown
      message: Waiting for agent initialization to finish
      since: 16 Jan 2016 00:28:39+07:00
    relations:
      db:
      - postgresql
      db-admin:
      - postgresql
  postgresql:
    charm: local:trusty/postgresql-11
    exposed: false
    life: dying
    service-status:
      current: error
      message: 'hook failed: "db-relation-changed"'
      since: 16 Jan 2016 00:48:55+07:00
    relations:
      db:
      - client
      db-admin:
      - client
      replication:
      - postgresql
    units:
      postgresql/17:
        workload-status:
          current: error
          message: 'hook failed: "db-relation-changed" for client:db'
          since: 16 Jan 2016 00:48:55+07:00
        agent-status:
          current: idle
          since: 16 Jan 2016 00:48:55+07:00
          version: 1.26-alpha3.1
        agent-state: error
        agent-state-info: 'hook failed: "db-relation-changed" for client:db'
        agent-version: 1.26-alpha3.1
        life: dying
        machine: "26"
        open-ports:
        - 5432/tcp

[Stuart Bishop (stub)]

machine-0.log, which is rather large. Note that this is the same environment as Bug #1532130 , which had occurred a short while earlier. There is a chance that the weirdness here is fallout from this earlier failure.

Also amusingly, lxc shows the machine aparently active with an IP address. Alas, processes kicked in and it got shutdown before I could investigate further or rescue the unit log from it.

$ lxc list
+------------------------------------------------------+---------+-------------------+------+-----------+-----------+
| NAME | STATE | IPV4 | IPV6 | EPHEMERAL | SNAPSHOTS |
+------------------------------------------------------+---------+-------------------+------+-----------+-----------+
| juju-5f65e7eb-5d0b-4c6e-8fde-143ad97856fc-machine-0 | RUNNING | 10.0.3.145 (eth0) | | NO | 0 |
+------------------------------------------------------+---------+-------------------+------+-----------+-----------+
| juju-5f65e7eb-5d0b-4c6e-8fde-143ad97856fc-machine-26 | RUNNING | 10.0.3.36 (eth0) | | NO | 0 |
+------------------------------------------------------+---------+-------------------+------+-----------+-----------+
| juju-81816a86-9dd3-445a-8bc0-99ff62d179c9-machine-0 | RUNNING | 10.0.3.144 (eth0) | | NO | 0 |
+------------------------------------------------------+---------+-------------------+------+-----------+-----------+

(Hmm... and a second controller node that didn't get cleaned up last teardown)

[Cheryl Jennings (cherylj)]

Stuart - are you still running into this issue?

If you are able to recreate, please grab the contents of /var/log/juju/* for the instance with the agent-state stuck in 'pending'. You should be able to ssh to the machine to get it, even if the agents haven't started

[Stuart Bishop (stub)]

I've never been able to reproduce the situation.

[Launchpad Janitor (janitor)]

[Expired for juju-core because there has been no activity for 60 days.]