Fix Validation using old tvars fails for Cylc 7 workflows

[wxtim]

Allow parsec to fail gracefully when trying to retrieve template variables from a workflow database if it's a Cylc 7 DB, by doing the same DB compatibility check that's done on restart.

Fixes two small issues not recorded elsewhere

• I had failed to use the is_public kwarg on CylcWorkflowDAO.__init__ , without which the function tries to write to the database, which we don't want to do for the parsec.fileparse use of database access. (Now sorted out by DB interface & safety improvements #5330)
• Pre Cylc 8 workflows don't have the database field under the same name - have made cylc play and cylc vr fail gracefully when trying to run on such workflows.

Also closes #5317

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.
• Tests are included (or explain why tests are not needed).
• CHANGES.md entry included if this is a change that can affect users
• Cylc-Doc pull request opened if required at cylc/cylc-doc/pull/XXXX.
• If this is a bug fix, PR should be raised against the relevant ?.?.x branch.

[oliver-sanders]

The check which prevents Cylc 8 from restarting Cylc 7 workflows is this:

Traceback (most recent call last):
  ...
  File "~/cylc-flow/cylc/flow/workflow_db_mgr.py", line 799, in check_workflow_db_compatibility
cylc.flow.exceptions.ServiceFileError: Workflow database is incompatible with Cylc 8.2.0.dev, or is corrupted.

So I think all that's needed is to add the check_workflow_db_compatibility check at the start of cylc vr.

BUT

Only do this if the workflow is NOT running (because if it is running, we know it's Cylc 8 compatible). Note it doesn't matter that check_workflow_db_compatibility uses the private database provided we know the workflow is not running.

[MetRonnie]

cylc-flow/cylc/flow/rundb.py

Lines 338 to 339 in 06b4079

	if not self.is_public:
	self.create_tables()

IMO whether to create the tables or not should be handled by a different, new arg, called create_tables instead of is_public, because that really isn't clear otherwise

[datamel]

I have checked out this branch. I have run workflows under cylc7 and then tried validate with this branch and master.
I get a nice error on this branch when the workflow has run and validating at cylc 8, master will state the workflow is valid at cylc 8.
If I validate an actively running (with cylc 7) workflow on this branch, it still pass validation. This is probably a niche case? EDIT: I see Oliver has this covered and it it ok in this case.
When I play an actively running cylc 7 workflow, I get traceback with KeyError: 'CYLC_WORKFLOW_PID' at cylc8, again probs niche and maybe a separate issue?

[wxtim]

IMO whether to create the tables or not should be handled by a different, new arg, called create_tables instead of is_public, because that really isn't clear otherwise

It might well have saved me from this error, but unless you insist I think it's out of scope.

[oliver-sanders]

Could you rebase onto 8.1.x please.

[oliver-sanders]

Ran through the methodology from here - #5313 (comment)

2023-01-24T10:10:23Z CRITICAL - Cannot tell if the workflow is running
    Note, Cylc 8 cannot restart Cylc 7 workflows.
Traceback (most recent call last):
  File "/home/h06/osanders/mambaforge/envs/cylc.dev/bin/cylc", line 33, in <module>
    sys.exit(load_entry_point('cylc-flow', 'console_scripts', 'cylc')())
  File "/net/home/h06/osanders/cylc-flow/cylc/flow/scripts/cylc.py", line 653, in main
    execute_cmd(command, *cmd_args)
  File "/net/home/h06/osanders/cylc-flow/cylc/flow/scripts/cylc.py", line 286, in execute_cmd
    entry_point.resolve()(*args)
  File "/net/home/h06/osanders/cylc-flow/cylc/flow/terminal.py", line 232, in wrapper
    wrapped_function(*wrapped_args, **wrapped_kwargs)
  File "/net/home/h06/osanders/cylc-flow/cylc/flow/scripts/validate_reinstall.py", line 120, in main
    sys.exit(vro_cli(parser, options, workflow_id))
  File "/net/home/h06/osanders/cylc-flow/cylc/flow/scripts/validate_reinstall.py", line 152, in vro_cli
    workflow_running, options.templatevars, options.templatevars_file
UnboundLocalError: local variable 'workflow_running' referenced before assignment

Need to halt the execution either by raising the original exception as suggested, or by return 1 or sys.exit(1) or whatnot. So long as the commands exit code is non-zero.

[MetRonnie]

Added some tests in wxtim#52

[oliver-sanders]

Re-reviewed code, LGTM, moving onto functional testing...

[MetRonnie]

Regression in latest rebase. Would not recommend rebasing at this point due to how out of date the earlier commits are, just whack squash 'n merge at the end

Update: I have force pushed back to f26b78c

[MetRonnie]

Have tested for both a running and stopped Cylc 7 workflow:

• cylc play
• cylc validate
• cylc vr (Produces traceback for a running Cylc 7 workflow but the correct error message is now printed at the top so who cares)
• cylc vip

And ensured no tables are created by these commands

[datamel]

Tested as working. Thanks @wxtim