setting max parallel running jobs

[buddhabas]

Hi all,
Is there a way to set maximum parallel job on queue?
If I start my transcoding there are 20 parallel jobs. I'd like to have max 5 for example
Thanks!

[emcodem]

Its a little confusing question, usually users want to set max paralell jobs running, i guess you found that setting already.
If you really want to define the max jobs in queue, you can do this actually but of course only related to watchfolder operations. API and manual jobs will just queue in normally (what else should they do).
In ffastrans.json find max_queue value. Also on webint ffastrans->Config menu, "Debug" tab.

[buddhabas]

Its a little confusing question, usually users want to set max paralell jobs running, i guess you found that setting already.
If you really want to define the max jobs in queue, you can do this actually but of course only related to watchfolder operations. API and manual jobs will just queue in normally (what else should they do).
In ffastrans.json find max_queue value. Also on webint ffastrans->Config menu, "Debug" tab.

Thanks forr feedback.
I'd like to have a user_variable to set on Populate_variables
Something like "max_parallel_running_jobs"

[emcodem]

Sorry but still not clear what you mean.

You can potentially alter the max_queue setting i mention above from a workflow but i believe this is not what you want.
So if there was a user_variable as you call it max_parallel_running_jobs, what exactly would it do?

[buddhabas]

Sorry but still not clear what you mean.

You can potentially alter the max_queue setting i mention above from a workflow but i believe this is not what you want.
So if there was a user_variable as you call it max_parallel_running_jobs, what exactly would it do?

If a start one of my workflows (I put in 20 files) I can see on monitor all 20 files running parallelement.
I would like to limit to 5

[emcodem]

and why you want to change this value via a "user_variable"?

[buddhabas]

and why you want to change this value via a "user_variable"?

I think it’s more convenient than edit a json.

[emcodem]

Understood that but does the max_queue setting in ffastrans.json actually do what you want?

[buddhabas]

I closed FFASTrans.
Set

Code: Select all

    "max_queue": 5

on ffastrans.json

Started ffastrans.
Manuallky submitted 20+ files to a WF.
I can see all job started, not only 5 as set.

Why?

[FranceBB]

are the 20 parallel jobs created because you're splitting the workflow in different sections all being executed at the same time like what you're doing with the audio extractor + H.264 video encoder on the other topic?
If you don't want them to be executed in parallel you can make them sequential, all you need is to access the original file, so you can either populate s_source with s_original_full and start back from that source file, then a new A/V Decoder etc etc etc or you could save the indexed source by putting a populate variable node after the first A/V Decoder to set s_my_source = s_source , and then after you've done your thing and delivered the first file, reference back to that with a populate variable to set back s_source to s_my_source.

I can send you a workflow as an example if you want but what I mean is:

watchfolder
Encoding node
delivery node (output 1)
populate node (s_source = s_original_full)
Encoding node
delivery node (output 2)

etc