Add support for sequential job definitions

[njakobsen]

Jobs now support a sequence option that will cause jobs with the same sequence value to run sequentially when scheduled for the same time. sequential: true can be passed to the every block to automatically set a common sequence value for each contained job.

every 3.hours, sequential: true do
  # Jobs in this block that don't set their own sequence will default to run in the same sequence (not in parallel)
  runner "MyModel.some_process", halt_sequence_on_failure: true # If this job fails, subsequent jobs will not run (defaults to false)
  rake "my:rake:task"
  command "/usr/bin/my_great_command", sequence: nil # This job runs in parallel with other jobs that are not part of a sequence
  runner "MyModel.task_to_run_at_four_thirty_in_the_morning", sequence: 'other' # This job runs sequentially with other jobs in the sequence called 'other'
end

Outstanding Questions

• Is there a need to configure the command delimiter? i.e. is && always what we want, or do we want to be able to use ; so the sequence continues even if one command fails. This feature has been added.

Note

• This PR also includes a previous PR of mine that allows the options set on an every block to be overridden by those set on the individual jobs in the block. That other PR has been merged into javan/main.

[taketo1113]

I think it might be better to use a name like chain or combine instead of sequence, to better reflect the actual behavior of the generated crontab. What do you think?

---

Here is the behavior I confirmed:

every 2.hours do
  command "blahblah", sequence: 'backups'
  command "foofoo", sequence: 'backups'
  command "barbar"
end

# crontab
0 0,2,4,6,8,10,12,14,16,18,20,22 * * * /bin/bash -l -c 'blahblah' ; /bin/bash -l -c 'foofoo'
0 0,2,4,6,8,10,12,14,16,18,20,22 * * * /bin/bash -l -c 'barbar'