Juju commands return SSL errors
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
pyjuju |
Invalid
|
High
|
Unassigned |
Bug Description
While waiting for a service to start our scripts must repeatedly query 'juju status' waiting for the agent-state to transition to 'started'. Infrequently but in a repeatable manner, 'juju status' will encounter an SSL error and fail.
Here is the output of a demonstration script I wrote that calls 'juju status' repeatedly, sleeping 0.1 seconds between calls:
/home/bac> ./status_fail.py
.......
Exception caught on iteration 133
Command '['juju', 'status']' returned non-zero exit status 1
2012-04-18 11:34:26,993 WARNING ssl-hostname-
2012-04-18 11:34:26,993 WARNING EC2 API calls encrypted but not authenticated
2012-04-18 11:34:26,993 WARNING S3 API calls encrypted but not authenticated
2012-04-18 11:34:26,993 WARNING Ubuntu Cloud Image lookups encrypted but not authenticated
2012-04-18 11:34:26,994 INFO Connecting to environment...
2012-04-18 11:34:28,317 INFO Connected to environment.
Connection to the other side was lost in a non-clean fashion: Connection lost.
2012-04-18 11:34:28,874 ERROR Connection to the other side was lost in a non-clean fashion: Connection lost.
Traceback (most recent call last):
File "./status_fail.py", line 17, in <module>
juju_status()
File "/usr/lib/
return run(*all_args)
CalledProcessError: Command '['juju', 'status']' returned non-zero exit status 1
(The script used can be seen at http://
Any way you can make your script use 'juju -v status' and give us the stderr?