Comment 8 for bug 1255242

Revision history for this message
Curtis Hovey (sinzui) wrote : Re: upgrade-juju and bootstrap on HP cloud broken in devel

Now that the bootstrap private ip fix is in trunk. CI now sees this bug when deploying wordpress and mysql using the the 1.17.0 candidate on both the client and server. CI sees a hook failure. 1.16.5 does not encounter this problem. This issue could relate to HP's speed or internals. The mysql charm itself might be brittle. I have played out this scenario 3 times using my one command line to allow for human slowness
    http://bazaar.launchpad.net/~juju-qa/juju-core/ci-cd-scripts2/view/head:/deploy_stack.py

We can see from the logs that the right juju version is always found so the script does not attempt to force the version.
I see config-changed hook failures of the mysql agent is stuck in pending for 40 minutes.

I am attaching a log from CI. I have confirmed using debug log that my own tests see this, but I think it is orthogonal:

machine-1: 2013-12-18 17:40:59 ERROR juju.worker.authenticationworker worker.go:68 adding current Juju keys to ssh authorised keys: generating key fingerprint: exit status 1
machine-1: 2013-12-18 17:40:59 ERROR juju runner.go:220 worker: exited "authenticationworker": adding current Juju keys to ssh authorised keys: generating key fingerprint: exit status 1
machine-1: 2013-12-18 17:40:59 INFO juju runner.go:254 worker: restarting "authenticationworker" in 3s
machine-1: 2013-12-18 17:41:02 INFO juju runner.go:262 worker: start "authenticationworker"
machine-1: 2013-12-18 17:41:02 WARNING juju.ssh authorisedkeys.go:244 invalid Juju ssh key : generating key fingerprint: exit status 1

The hook failures seen in 2 of my 3 manual tests report
ubuntu0.12.04.1) ...
2013-12-18 17:55:51 INFO juju.worker.uniter context.go:323 HOOK 131218 17:55:51 [Warning] Using unique option prefix key_buffer instead of key_buffer_size is deprecated and will be removed in a future release. Please use the full name instead.
2013-12-18 17:56:00 INFO juju.worker.uniter context.go:323 HOOK mysql start/running, process 20579
2013-12-18 17:56:00 INFO juju.worker.uniter context.go:323 HOOK Setting up libhtml-template-perl (2.10-1) ...
2013-12-18 17:56:00 INFO juju.worker.uniter context.go:323 HOOK Setting up mysql-server (5.5.34-0ubuntu0.12.04.1) ...
2013-12-18 17:56:01 INFO worker.uniter.jujuc server.go:102 running hook tool "juju-log" ["-l" "INFO" "dataset size in bytes: 822083584"]
2013-12-18 17:56:01 INFO juju juju-log.go:66 mysql/0: dataset size in bytes: 822083584
2013-12-18 17:56:03 INFO juju.worker.uniter context.go:323 HOOK mysql stop/waiting
2013-12-18 17:56:05 INFO juju.worker.uniter context.go:323 HOOK start: Job failed to start
2013-12-18 17:56:05 INFO juju.worker.uniter context.go:323 HOOK Traceback (most recent call last):
2013-12-18 17:56:05 INFO juju.worker.uniter context.go:323 HOOK File "/var/lib/juju/agents/unit-mysql-0/charm/hooks/config-changed", line 356, in <module>
2013-12-18 17:56:05 INFO juju.worker.uniter context.go:323 HOOK check_call(['service','mysql','start'])
2013-12-18 17:56:05 INFO juju.worker.uniter context.go:323 HOOK File "/usr/lib/python2.7/subprocess.py", line 511, in check_call
2013-12-18 17:56:05 INFO juju.worker.uniter context.go:323 HOOK raise CalledProcessError(retcode, cmd)
2013-12-18 17:56:05 INFO juju.worker.uniter context.go:323 HOOK subprocess.CalledProcessError: Command '['service', 'mysql', 'start']' returned non-zero exit status 1