LXC container failed to start on 12.04

[Haitao Jiang]

Thanks.

I would appreciate if anyone can help me with the unit.log not being created.

Looking at container.log, the last few lines
:
      lxc-start 1336514321.869 NOTICE   lxc_conf -
'haitaoj-local-mysql-0' is setup.
      lxc-start 1336514321.869 INFO     lxc_start - changed apparmor
profile to lxc-container-default
      lxc-start 1336514321.869 NOTICE   lxc_start - exec'ing '/sbin/init'
      lxc-start 1336514321.869 NOTICE   lxc_start - '/sbin/init'
started with pid '11565'
      lxc-start 1336514321.869 WARN     lxc_console - console input disabled
      lxc-start 1336514321.869 WARN     lxc_start - invalid pid for SIGCHLD

mysql/0 status is still pending

Thanks

Haitao

On Tue, May 8, 2012 at 2:39 PM, Serge E. Hallyn <serge at hallyn.com> wrote:
> Quoting Haitao Jiang (jianghaitao at gmail.com):
>> Serge,
>>
>> Thanks. Let me put in more detail - sorry for the long email.
>>
>> I think my previous problems was due to that I have another juju
>> installation from the source code, thus messed things up. Now, I
>> deleted that and did following.
>>
>> 1. I upgraded juju etc. to the newest version. For juju, it is at
>> 0.5+bzr531-0ubuntu1
>> $ which juju
>> /usr/bin/juju
>>
>> 2. $ more .juju/environments.yaml
>> default: local
>> environments:
>>   local:
>>     type: local
>>     admin-secret: whatever
>>     data-dir: /home/haitaoj/lxc
>>     default-series: precise
>>
>> 3. Bootstrap
>> $ juju bootstrap
>> Checked the ~/lxc/haitaoj-local/machine-agent.log, machine agent
>> starts o, no error
>>
>> 4
>> $ juju status
>> machines:
>>   0:
>>     dns-name: localhost
>>     instance-id: local
>>     instance-state: running
>>     state: running
>> services: {}
>> 2012-05-08 11:03:10,590 INFO 'status' command finished successfully
>>
>> 5.
>> $ juju deploy ~/
>>
>> Now, machine-agent.log seems ok except the last line.
>> --------
>> :
>> 2012-05-08 11:03:39,982: juju.agents.machine at DEBUG: Units changed
>> old:set([]) new:set(['mysql/0'])
>> 2012-05-08 11:03:39,983: juju.agents.machine at DEBUG: Starting service
>> unit: mysql/0 ...
>> 2012-05-08 11:03:40,023: juju.agents.machine at DEBUG: Downloading charm
>> local:precise/mysql-122 to /home/haitaoj/lxc/haitaoj-local/charms
>> 2012-05-08 11:03:40,094: juju.agents.machine at DEBUG: Starting service
>> unit mysql/0
>> 2012-05-08 11:03:40,094: unit.deploy at DEBUG: Creating master container...
>> 2012-05-08 11:05:02,002: unit.deploy at DEBUG: Created master container
>> haitaoj-local-0-template
>> 2012-05-08 11:05:02,002: unit.deploy at INFO: Creating container mysql-0...
>> 2012-05-08 11:05:04,701: unit.deploy at INFO: Container created for mysql/0
>> 2012-05-08 11:05:04,735: unit.deploy at DEBUG: Charm extracted into container
>> 2012-05-08 11:05:04,736: unit.deploy at DEBUG: Starting container...
>> 2012-05-08 11:05:05,871: unit.deploy at INFO: Started container for mysql/0
>> 2012-05-08 11:05:05,871: juju.agents.machine at INFO: Started service unit mysql/0
>> 2012-05-08 11:05:51,665:11237(0x7feb6bdb8700):ZOO_WARN at zookeeper_interest@1461:
>> Exceeded deadline by 1528ms
>> ------------
>>
>> Things looks ok in units/master-customize.log
>>
>> units/mysql-0/container.log looks ok too.
>>
>> BUT, units/mysql-0/unit.log is not created and service is still in
>> pending state.
>>
>> So my question now is where the deploy process got stuck? Why unit.log
>> was not created?
>
> Looks like you've gotten past all the container related stuff, so I
> should probably defer to someone else - but just to check, at this point,
> if you do 'lxc-list', do you see a mysql container, and is it running,
> and if not does 'sudo lxc-start -n <thatcontainername>' work?
>
>> Thanks a lot for your help
>>
>> Haitao