Manual provisioning - feedback wanted

Thu Sep 26 01:45:14 UTC 2013

Matt,

Thanks for doing some more testing. So, short story is, manual provisioning
needs some more work to work with existing providers. The manually
provisioned node thinks it's running an ec2-provider instance. Thus, it
tries to do ec2isms, like querying 169.254.169.254 for instance metadata
(which will fail outside of the EC2 environment, as you can see here).

I think it's probably prudent we hold off on documenting/announcing this
wider until the null provider is available, or more work is done to enable
this sort of scenario (if that's even sensible/useful in a real world
deployment?)

Cheers,
Andrew

On Wed, Sep 25, 2013 at 10:53 PM, Matthew Williams <
matthew.williams at canonical.com> wrote:

> Hi Andrew,
>
> I've just tried this again on tip. I tried two separate providers that I'm
> signed up to. I tried to deploy the mongodb charm. Both seem to fail
> getting the private address? It happened on the wordpress charm as well so
> I guess it's juju related?
>
> Deploying mongodb on digital ocean:
>
> 2013-09-25 14:27:06 DEBUG juju.agent agent.go:241 Reading agent config,
> format: format 1.16
> 2013-09-25 14:27:06 INFO juju.jujud unit.go:73 unit agent unit-mongodb-0
> start
> 2013-09-25 14:27:06 INFO juju runner.go:253 worker: start "api"
> 2013-09-25 14:27:06 INFO juju apiclient.go:111 state/api: dialing "wss://
> ec2-23-20-9-138.compute-1.amazonaws.com:17070/"
> 2013-09-25 14:27:06 INFO juju apiclient.go:121 state/api: connection
> established
> 2013-09-25 14:27:06 INFO juju runner.go:253 worker: start "upgrader"
> 2013-09-25 14:27:06 INFO juju runner.go:253 worker: start "logger"
> 2013-09-25 14:27:06 DEBUG juju.worker.logger logger.go:35 initial log
> config: "<root>=DEBUG"
> 2013-09-25 14:27:06 INFO juju runner.go:253 worker: start "uniter"
> 2013-09-25 14:27:06 DEBUG juju.worker.logger logger.go:60 logger setup
> 2013-09-25 14:27:06 DEBUG juju.worker.logger logger.go:45 reconfiguring
> logging from "<root>=DEBUG" to "<root>=WARNING"
> 2013-09-25 14:28:09 ERROR juju.worker.uniter modes.go:41 cannot get unit's
> private address: cannot get "
> http://169.254.169.254/2011-01-01/meta-data/local-hostname": Get
> http://169.254.169.254/2011-01-01/meta-data/local-hostname: dial tcp
> 169.254.169.254:80: connection timed out
> 2013-09-25 14:28:09 ERROR juju.worker.uniter.filter filter.go:117 tomb:
> dying
>
> Deploying mongodb in greenqloud:
>
> 2013-09-25 14:46:51 INFO juju.jujud unit.go:73 unit agent unit-mongodb-0
> start
> 2013-09-25 14:46:51 INFO juju runner.go:253 worker: start "api"
> 2013-09-25 14:46:51 INFO juju apiclient.go:111 state/api: dialing "wss://
> ec2-23-20-9-138.compute-1.amazonaws.com:17070/"
> 2013-09-25 14:46:52 INFO juju apiclient.go:121 state/api: connection
> established
> 2013-09-25 14:46:52 INFO juju runner.go:253 worker: start "upgrader"
> 2013-09-25 14:46:52 INFO juju runner.go:253 worker: start "logger"
> 2013-09-25 14:46:52 DEBUG juju.worker.logger logger.go:35 initial log
> config: "<root>=DEBUG"
> 2013-09-25 14:46:52 INFO juju runner.go:253 worker: start "uniter"
> 2013-09-25 14:46:52 DEBUG juju.worker.logger logger.go:60 logger setup
> 2013-09-25 14:46:53 DEBUG juju.worker.logger logger.go:45 reconfiguring
> logging from "<root>=DEBUG" to "<root>=WARNING"
> 2013-09-25 14:46:58 ERROR juju.worker.uniter modes.go:41 cannot get unit's
> private address: cannot get "
> http://169.254.169.254/2011-01-01/meta-data/local-hostname": bad http
> response 404 Not Found
> 2013-09-25 14:46:58 ERROR juju.worker.uniter.filter filter.go:117 tomb:
> dying
>
> Thanks
>
> Matt
>
>
> On Wed, Sep 18, 2013 at 2:21 AM, Andrew Wilkins <
> andrew.wilkins at canonical.com> wrote:
>
>> On Wed, Sep 18, 2013 at 12:46 AM, Matthew Williams <
>> matthew.williams at canonical.com> wrote:
>>
>>> Hi Andy,
>>>
>>> When you say
>>> "Currently you do need to have an existing, bootstrapped environment.
>>> Work on improving this situation is underway"
>>>
>>> What's the idea?
>>>
>>
>> So there's another couple of things in development: a "null" provider,
>> and "manual bootstrapping". The null provider is a provider that does not
>> manage or provide any means of automatically provisioning machines. Manual
>> bootstrapping means essentially the same as manual provisioning, but for
>> the bootstrap node. In its current manifestation (under review), manual
>> bootstrap only works with the null provider. You will configure an
>> environment.yaml that names the bootstrap machine, and "juju bootstrap"
>> will do something similar to "juju add-machine ssh:..." under the hood.
>>
>> I miswrote earlier: you'll always need an existing, bootstrapped
>> environment. But with the null provider/manual bootstrap, you can put
>> everything in your VPS, or what have you. You won't need to have a cloud
>> environment just for the bootstrap node.
>>
>> Cheers,
>> Andrew
>>
>> Matt
>>>
>>>
>>> On Mon, Sep 9, 2013 at 2:47 AM, Andrew Wilkins <
>>> andrew.wilkins at canonical.com> wrote:
>>>
>>>> Hello everyone,
>>>>
>>>> As was noted in the 1.13.3 release notes thread, we did not announce
>>>> one the major features that made it into the release (manual provisioning).
>>>> This was intentional as we have not written any documentation yet. On the
>>>> other hand, it would be good to get some feedback so that we can make
>>>> changes if necessary, or feed into the documentation.
>>>>
>>>> So, below is a bit of a run-down. If you're interested in this feature,
>>>> try it out and let us know if you or have any issues or any thoughts for
>>>> improvement.
>>>>
>>>> ----
>>>>
>>>> As of 1.13.3 you can now do this:
>>>>     juju add-machine ssh:[user@]host
>>>>
>>>> and a series of commands will be carried out on that host, via SSH, to
>>>> provision the machine into the current juju environment. This will enable
>>>> you to compose a juju environment out of your existing systems.
>>>>
>>>> Here's a few things to bear in mind:
>>>>  - Currently you do need to have an existing, bootstrapped environment.
>>>> Work on improving this situation is underway
>>>>  - The machine you're provisioning must be able to route to machine 0
>>>> (for the state/API), and storage (to get tools, etc.)
>>>>  - There is no change in supported operating systems; the machine being
>>>> provisioned must be running Ubuntu 12.04+
>>>>  - Multiple invocations of ssh will be made, and sudo is used on the
>>>> remote host to install the machine agent. To reduce noisy prompts, you
>>>> should use public key authentication. To completely eliminate prompting,
>>>> you'll also need to enable passwordless sudo on the target host.
>>>>
>>>> Cheers,
>>>> Andrew
>>>>
>>>> --
>>>> Juju mailing list
>>>> Juju at lists.ubuntu.com
>>>> Modify settings or unsubscribe at:
>>>> https://lists.ubuntu.com/mailman/listinfo/juju
>>>>
>>>>
>>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.ubuntu.com/archives/juju/attachments/20130926/23c05acc/attachment-0001.html>