add platform on reload for SUCCEEDED and FAILED tasks

[wxtim]

These changes close #5063

Requirements check-list

• I have read CONTRIBUTING.md and added my name as a Code Contributor.
• Contains logically grouped changes (else tidy your branch by rebase).
• Does not contain off-topic changes (use other PRs for other changes).
• Applied any dependency changes to both setup.cfg and conda-environment.yml.
• Appropriate tests are included (unit and/or functional).
• Appropriate change log entry included.
• No documentation update required.
• Merged bugfix onto master

[oliver-sanders]

(happy omitting tests for this one, we need to review polling logic at some point and develop a complete set of unit tests)

[MetRonnie]

LGTM

[oliver-sanders]

Changed base to 8.0.x to match the milestone.

Could we have a changelog entry for this and a more informative commit message than "attempt to fix".

[oliver-sanders]

Tried this with Dave's example from #4513, I think I've managed to reproduce it on this branch. Will investigate and try again, could someone else have a crack.

[wxtim]

Tried this with Dave's example from #4513, I think I've managed to reproduce it on this branch. Will investigate and try again, could someone else have a crack.

Actually closes #5063/

[MetRonnie]

I think #5063 is a duplicate of #4513, which is the 8.0.1 version of #4516

Ah wait: #4516 (comment)

[MetRonnie]

I tried reproducing #4516 on 8.0.x, but don't think I could reproduce using _remote_background_indep_tcp platform

[scheduling]
    [[graph]]
        R1 = a:fail? => restart

[runtime]
    [[a]]
        script = sleep 60
        execution time limit = PT1S
        platform = _remote_background_indep_tcp
    [[restart]]
        script = """
            cylc stop "${CYLC_WORKFLOW_ID}" --now --now
            sleep 5
            cylc play "${CYLC_WORKFLOW_ID}" --host=localhost
        """

[wxtim]

I tried reproducing #4516 on 8.0.x, but don't think I could reproduce using _remote_background_indep_tcp platform

[scheduling]
    [[graph]]
        R1 = a:fail? => restart

[runtime]
    [[a]]
        script = sleep 60
        execution time limit = PT1S
        platform = _remote_background_indep_tcp
    [[restart]]
        script = """
            cylc stop "${CYLC_WORKFLOW_ID}" --now --now
            sleep 5
            cylc play "${CYLC_WORKFLOW_ID}" --host=localhost
        """

TBH I did the testing manually - I couldn't get this auto restarter working well and used @dpmatthews 's example.

[wxtim]

Example I used

#!jinja2
[scheduling]
    [[graph]]
        R1 = "timeout-remote & timeout-local-slurm & timeout-local-bg"
[runtime]
    [[timeout-remote]]
        script = "sleep 600"
        platform = xce
        execution time limit = PT1S
    [[timeout-local-slurm]]
        script = "sleep 600"
        platform = spice
        execution time limit = PT1S
    [[timeout-local-bg]]
        script = "sleep 600"
        execution time limit = PT1S

Wait for all three to fail, then stop and restart the workflow.

[oliver-sanders]

add platform on reload for SUCCEEDED and FAILED tasks

This issue is about "restart" rather than "reload" right?

[oliver-sanders]

Managed to reproduce with the timeout-remote platform in Tim's example on master, couldn't reproduce on this branch.

LGTM.

[oliver-sanders]

@wxtim could you add a changelog.

[wxtim]

I've rebased this on the wrong branch. I'm taking that as an indication of tiredness. Will fix tomorrow. 😞

[wxtim]

@oliver-sanders @MetRonnie Please check that I haven't messed owt up with my foolish rebasing.

[oliver-sanders]

(letting tests pass)