Hello,

thanks a lot for your answers, and the last one  :), is there any plan to keep on working with this feature?

Regards

On 6 February 2015 at 11:39, Peter Cock <p.j.a.cock@googlemail.com> wrote:
On Fri, Feb 6, 2015 at 9:59 AM, Roberto Alonso CIPF <ralonso@cipf.es> wrote:
> Hello,
>
> Postgres did it perfectly, but as I would like to enable this option in my
> production servers I changed two configs in galaxy.ini:
>
> # This enables splitting of jobs into tasks, if specified by the particular
> tool config.
> # This is a new feature and not recommended for production servers yet.
> use_tasked_jobs = True
> local_task_queue_workers = 4
>
> So I have two questions,
> 1. Why is this not appropriate for productions server yet?

At a practical level, I believe this setting is not used on the main
usegalaxy.org instance, and thus has not been stress tested by
the Galaxy developers.

Also (see below), it has little documentation.

> 2. What is local_task_queue_workers doing?
>

I'm not entiruely sure, but I think it is how many task-splitters
would be run at once (e.g. several splittable jobs are submitted
together). We use local_task_queue_workers = 4 but I forget
why exactly.

> I think I misunderstand something
> because I put this parameter in the tool config file:
>    <parallelism method="basic" split_size="20"
> split_mode="number_of_parts"></parallelism>
> on ther other hand, I don't see any documentation about parallelism tag, is
> there any?

I don't think there is anything explicit on the wiki about how
to use the parallelism tag :(

Peter



--
Roberto Alonso
Functional Genomics Unit
Bioinformatics and Genomics Department
Prince Felipe Research Center (CIPF)
C./Eduardo Primo Yúfera (Científic), nº 3
(junto Oceanografico)
46012 Valencia, Spain
Tel: +34 963289680 Ext. 1021
Fax: +34 963289574
E-Mail: ralonso@cipf.es