Julia 1.2-1.6 does not do task switch. If multiple jobs were running on the same thread as job scheduler, or other tasks that need instant response, the heavy jobs would block job scheduler and other important tasks.

Solution: ThreadPools.jl allows allocate jobs at a specific thread. Try to implement it in JobScheduler.

Related to https://docs.julialang.org/en/v1/devdocs/precompile_hang/

I think it might be because of the scheduler_start call inside of __init__. Is it necessary to have the scheduler start when first loading the module?

I have a recurring job whose function can sometimes take longer to finish than the actual period of recurrence. The subsequent call of function would be independent of the previous one, however due to the implicit dependence, it's not started until the previous one is finished. Would it be possible to add some option that removes this dependence?

As an MWE, this is what I have now (the effective function call period becomes 6 seconds):

julia> using JobSchedulers
julia> function f()  # function runs within 4 seconds
           sleep(4)
           println(Dates.format(now(), "HH:MM:SS"))
       end;
julia> submit!(f, cron=Cron("*/2", "*", "*", "*", "*", "*"));  # function call at every 2 second, starting at 15:19:16
15:19:20
15:19:26
15:19:32
15:19:38
...

What would be great to have is something like this (the function call is 2 seconds as specified by cron):

 julia> submit!(f, cron=Cron("*/2", "*", "*", "*", "*", "*"), remove_self_dependence=true);  # function call at every 2 second, starting at 15:19:16
15:19:20
15:19:22
15:19:24
15:19:26
...

1. Only show progress meter when stdout is a TTY. If stdout is not a TTY, only show the latest result with plain text.
2. Add ESC to stop progress meter when isinteractive().

How to pass parameter for a job ? My job is recurring I mean a Task must run every 3 minutes. Please guide me.

Hello,

The following code

using JobSchedulers

recurring_job = Job(cron = Cron(0, "*/5", *, *, *, *)) do
    println("function job done at $(now())")
end

submit!(recurring_job)

while true
    sleep(0.01)
end

run like this

>julia sched.jl
function job done at 2023-07-13T10:09:32.344
function job done at 2023-07-13T10:10:00.029
function job done at 2023-07-13T10:15:00.020
function job done at 2023-07-13T10:20:00.056

Is there a way to avoid first run ie ensure that first run is done at second=0 and minute as multiple of 5 ?

Kind regards

This issue is used to trigger TagBot; feel free to unsubscribe.

If you haven't already, you should update your TagBot.yml to include issue comment triggers.
Please see this post on Discourse for instructions and more details.

If you'd like for me to do this for you, comment TagBot fix on this issue.
I'll open a PR within a few hours, please be patient!

Not sure if this is a feature request or a question: I'm not sure how much memory my Jobs take relative to the inputs they receive. I'd like to update the estimate I provide to the mem keyword based on the amount of memory the last jobs took. Is the high-water mark of memory consumed by a given job recorded by the scheduler? How could I access it?

If it isn't recorded, would it be feasible to add that feature?

1. dependency = DONE => job_A: to be simplified to dependency = job_A or dependency = job_A.id

2. submit!(Job(...)): to be simplified to submit!(...)

3. Set a lower loop interval.

4. Move scheduler_start() in __init__.