[FEATURE] jobify ir_cron

[Cedric-Pigeon]

Hello,

The most painpoint with ir_cron is that you are not able to have a feedback in case of failure except by reading server logs.
My intend is to create a new module adding a flag "Manage by job" on a cron.
If this flag is checked, the user can specify a channel to use, default will be root, and each execution will be done through a queue job.
By this way users will be able to track faillings cron and eventually relaunch it.

Any feedback on that ?

[Cedric-Pigeon]

@guewen @lmignon

[sbidoul]

I'd call the flag "Run as Queue Job". Otherwise 👍 great idea.

queue_job_cron as module name?

[JordiBForgeFlow]

That would be great!! 👍

[Cedric-Pigeon]

queue_job_cron seems good to me.

[lmignon]

👍 great Idea!

[simahawk]

👍

[guewen]

I like this idea a lot.

A random thought: it could maybe allow several types of time intervals : N minutes after the end of the last execution, N minutes after the last execution except if still running, N minutes after the last execution no matter what and allowing overlap (need the method to be compatible).

[JordiBForgeFlow]

Would be also very good to establish dependencies.between jobs to prevent potentially conflicting different jobs to overlap. See OCA/server-tools#999

…

On Dec 1, 2017 8:00 PM, "Guewen Baconnier" ***@***.***> wrote: I like this idea *a lot*. A random thought: it could maybe allow several types of time intervals : N minutes after the end of the last execution, N minutes after the last execution except if still running, N minutes after the last execution no matter what and allowing overlap (need the method to be compatible). — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#40 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AHU_VgDL0K85esmlvrnRxdSfPBS8qlV0ks5s8EzrgaJpZM4QyXzO> .

[sbidoul]

dependencies.between jobs

@jbeficent there is a concept of sequential channels in the queue job runner that guarantees job execution order. That should be sufficient for most purposes.

[sbidoul]

@guewen we already have is the throttling option in channels configuration.

[JordiBForgeFlow]

Oh, so even whe jobs are assigned to different ir.cron, if they belong to the same channel they would be executed sequentially, right? I can think, though, of jobs with different execution frequencies that you would not want to conflict. But perhaps this is too specific to a problem we face in DDMRP where we need.to update different attributes of the buffer in separate crons (Average Daily Usage is computed every 8 hrs and DDMRP alerts every 2 hrs).

…

On Dec 1, 2017 8:17 PM, "Stéphane Bidoul (ACSONE)" ***@***.***> wrote: dependencies.between jobs @jbeficent <https://github.com/jbeficent> there is a concept of sequential channels in the queue job runner that guarantees job execution order. That should be sufficient for most purposes. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#40 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AHU_VtWA15Rf9HxNx0bOiNYhSUf3BRbTks5s8FDbgaJpZM4QyXzO> .

[sbidoul]

@jbeficent actually you can simply put all these jobs in a channel with capacity 1 (not necessarily sequential) to be sure they will not run concurrently.

[JordiBForgeFlow]

Good! Lookong forward to using it!

[elicoidal]

👍