Firstly I would like to thank all the contributor of this project. I have been using it for a while and it’s working great.
I have couple of quick questions to ask which are
- I see that the primary key of the job table is
bigint. My question is that what if I created jobs so much that eventually ran out of thebigint’s valid range. Does Oban provide a mechanism to resolve this issue yet?. Currently I am using customer pruner to remove jobs from thedbalready but I would like to know if I ran out the validbigintto create a primary key what could happen? - currently I am using oban on the distributed nodes. I am having issues to set the cron, to uniquely run a single long running task across nodes, And I want to make sure that until a job is completed the new job should not be scheduled, I am using oban
cronjobmachanism to do this.
This is what my job module looks like →
# My Job Module
defmodule MyApp.Job do
use Oban.Worker,
queue: :worker1,
unique: [
fields: [:args],
states: [:available, :scheduled, :executing],
period: 60
]
def perform(_) do
end
end
# My cron configuration in config
config :my_app, Oban,
repo: My.Repo,
queues: [worker1: 1],
crontab: [
{"* * * * *", MyApp.Job}
]
With this configuration, I could see that new job is never scheduled until the current job is finished. But I am seeing that the new job is created before a time diff of 1 minute. One more thing to note is that this job could be running for small time or it could run for serveral minutes. I want to guarantee two things → 1. only 1 job can be run at a time on distributed nodes, and 2. I want to gurantee that a new job should not be created before 1 minute
Thanks
Suracheth
