Comment 5 for bug 1218095

Revision history for this message
Antonio Terceiro (terceiro) wrote : Re: [Bug 1218095] Re: lava-deployment-tool: line 286: install_devmode: command not found

On Thu, Aug 29, 2013 at 08:27:57PM -0000, Kim Phillips wrote:
> > For me it just works. I would try on Ubuntu 12.04 which is at the
> moment the only supported platform.
>
> excerpt from [1]:
>
> "Software Requirements
> We currently recommmend using Ubuntu 12.04. This tool should work on Ubuntu versions 11.10 (Oneiric) and newer.
> If you’d like to help us with other distributions feel free to contact us at linaro-validation (at) lists (dot) linaro (dot) org."
>
> I filed this bug as a response to the invitation in the last sentence,
> and to improve the odds behind the word 'should' in the second sentence.
>
> My current 'workaround' is to use the cambridge lab server, as -
> although slower - it's a better option than reverting my development
> machine's OS.

Sure, appreciated for trying.

The problem is, once you decide to use a platform where it "should" work
but is not supported, then you have to help us fix the problems. :-)

> > I confirm the problem with `install_devmode: command not found`, but
> it is innofensive, just annoying. I am pushing a fix for that.
>
> ok, thanks.
>
> I still have a problem getting a job to go from 'submitted' state to
> 'started'. Based on IRC discussions, the scheduler daemon is not
> starting / empty scheduler log file.
>
> Looking at the log again, could this be the problem?:
>
> FATAL ERROR - The following SQL query failed: CREATE LANGUAGE plpgsql
> The error was: language "plpgsql" already exists

I don't think so. I also get this error, and AFAICT it's also
innofensive.

So your problem is that the scheduler is not starting.
lava-deployment-tool will create upstart jobs for all of the components
that need to be started, maybe you can investigate whether there is some
problem with those upstart jobs?

Did you try `start lava` explicitly to see what happens?