LAVA Scheduler (deprecated)

Merge lp://staging/~doanac/lava-scheduler/celery-removal into lp://staging/lava-scheduler

celery-removal
Merge into trunk

Proposed by Andy Doan on 2012-12-03

Status:	Merged
Merged at revision:	230
Proposed branch:	lp://staging/~doanac/lava-scheduler/celery-removal
Merge into:	lp://staging/lava-scheduler
Diff against target:	392 lines (+178/-47) 6 files modified lava_scheduler_app/management/commands/schedulermonitor.py (+2/-5) lava_scheduler_app/migrations/0028_auto__del_field_devicetype_use_celery.py (+151/-0) lava_scheduler_app/models.py (+0/-7) lava_scheduler_daemon/board.py (+12/-24) lava_scheduler_daemon/dbjobsource.py (+11/-2) lava_scheduler_daemon/service.py (+2/-9)
To merge this branch:	bzr merge lp://staging/~doanac/lava-scheduler/celery-removal
Related bugs:	Link a bug report

Reviewer	Review Type	Date Requested	Status
Linaro Validation Team		2012-12-03	Pending
Review via email: mp+137476@code.staging.launchpad.net

Description of the change

building off the recent lava-dispather "devices" API, this allows us to run multiple scheduler instances. Each instance only performs jobs for devices configured by that instance's local lava-dispatcher devices.

with that in place, we really don't need celery so its been removed.

the key thing that will bring this all together is an update I'm doing in lava-deployment-tool that creates "worker nodes".

Revision history for this message

Michael Hudson-Doyle (mwhudson) wrote on 2012-12-03:

Seems OK. Do we need to do anything operational before we can
merge/deploy this? (Set up sshfs maybe?)

Revision history for this message

Andy Doan (doanac) wrote on 2012-12-03:

On 12/03/2012 01:29 PM, Michael Hudson-Doyle wrote:
> Seems OK. Do we need to do anything operational before we can
> merge/deploy this? (Set up sshfs maybe?)
>
no - thanks to you lava-deployment-tool's evolution support will allow
the next upgrade of control to set things up properly. Additionaly, I
think I'll just install from scratch on our worker nodes to keep things
nice and clean.

Revision history for this message

Michael Hudson-Doyle (mwhudson) wrote on 2012-12-03:

Andy Doan <email address hidden> writes:

> On 12/03/2012 01:29 PM, Michael Hudson-Doyle wrote:
>> Seems OK. Do we need to do anything operational before we can
>> merge/deploy this? (Set up sshfs maybe?)
>>
> no - thanks to you lava-deployment-tool's evolution support will allow
> the next upgrade of control to set things up properly. Additionaly, I
> think I'll just install from scratch on our worker nodes to keep things
> nice and clean.

OK. Merge when ready then (after today's deployment?)

Revision history for this message

Andy Doan (doanac) wrote on 2012-12-03:

On 12/03/2012 02:03 PM, Michael Hudson-Doyle wrote:
> Andy Doan <email address hidden> writes:
>
>> On 12/03/2012 01:29 PM, Michael Hudson-Doyle wrote:
>>> Seems OK. Do we need to do anything operational before we can
>>> merge/deploy this? (Set up sshfs maybe?)
>>>
>> no - thanks to you lava-deployment-tool's evolution support will allow
>> the next upgrade of control to set things up properly. Additionaly, I
>> think I'll just install from scratch on our worker nodes to keep things
>> nice and clean.
>
> OK. Merge when ready then (after today's deployment?)
>
yes - *after*. I'd like to do some more extensive testing :)