[ANN] juju-core 1.9.1 has been released

David Cheney david.cheney at canonical.com
Fri Nov 2 07:48:52 UTC 2012


> This seems like a bug. Warning and ignoring on an unimplemented provider
> would seem a better option than just refusing it. I don't see a downside
> other than a lot of warnings about unimplemented providers until they
> are implemented. :)

After talking to Jim Baker, I've raised 
https://bugs.launchpad.net/juju-core/+bug/1074025 which will be fixed in 
short order. If you can raise a bug if you have configurations that 
1.9.1 refuses to consume that would be great. 1.9.1 should ignore 
configurations from providers that
we dont yet support, but it may not be as well tested as it should be as 
we've been focused on EC2.

>> * remove-unit currently does not cause the underlying instance to shut
>> down. This somewhat reduces the usefulness of the previous workaround.
>> #1067127
>
> Isn't that intentional so we can decide to back up the data before
> terminating the machine?

Yes, after talking to William there are actually two issues, the first 
is there is no terminate-machine command, 
https://bugs.launchpad.net/juju-core/+bug/1073976, and the second is 
there are a few steps that the current unit agent is not taking to 
properly cleanup after itself (the service-unit reference stays around 
in status) but the instance is free'd to be reused by other units.

Cheers

Dave



More information about the Juju-dev mailing list