-
Notifications
You must be signed in to change notification settings - Fork 93
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
7.8.x Correctly store stop_point for restart #3147
Conversation
618f8ea
to
abd4e89
Compare
Note: We seem to be mixing initial cycle point with start point and final cycle point with stop point in many parts of the logic. As mentioned in #2799, we have an inconsistency between And the water gets even muddier with the |
Some of that confusion might pre-date our ISO8601 support, when those distinctions didn't really exist. |
03765f5
to
7dbda01
Compare
7dbda01
to
e03d129
Compare
e03d129
to
06e9f91
Compare
06e9f91
to
7ce08ae
Compare
7ce08ae
to
ed876ec
Compare
Store stop point in suite run time SQLite file. Slightly improved interface for telling restart to ignore previous initial, start, final and/or stop points.
Rationalise these settings between suite start up options and suite configuration, and prevent repetitive propagation of attributes. Store the user specified suite start up options as opposed to the calculated values - because those values are calculated by combining the original user specified value with the loaded configuration on restart.
Test setting stop point on suite start up and via cylc stop command. Test restart with stop point. Test stop with stop point consumed and restart. Test reload with stop point after restart.
ed876ec
to
633ce60
Compare
Will update non-local tests. (I have deleted restart/36 and can get restart/37 to work, but restart/34,41,42 are still failing for some reason. I'll look further on Monday.) |
c3ba8ba
to
eee751b
Compare
I now have a slightly better handle of why some of the non-local tests are failing. (Guess what?) It is to do with:
(I'll check that the tests fail in the same way under the vanilla 7.8.x branch or not on Monday.) |
This all sounds oddly familiar... |
On the failing tests:
|
Added change log. |
|
Have tested restart compatibility with Cylc 7.8.0, can confirm that CLI stop point is preserved with restart with this branch. +1 |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I'm happy with this, approved pending the doc change.
I'm not too fond of variables like holdcp
and stopcp
(perhaps stop_cp
) but once you know what they are it's OK.
Now that stop point, etc are stored correctly, the only thing that limits auto stop-restart is no detach mode.
Doc updated. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Nice work. I have not found any problems in a quick review+test ("time is of the essence", at the moment!); thanks to @oliver-sanders for the more thorough review.
Address mainly #2799 and #3019 (with consideration of #3161) for 7.8.x.
Changes:
suite.rc
, so it is safer to store the initial options rather than the values of the latest state.)cylc.config.SuiteConfig
class, by making use of the start up options (normally from CLI option parser) directly.--icp=CYCLE-POINT
forcylc run
so it has a more uniform interface with other commands, and with the final/stop cycle point options. Nowcylc run SUITE POINT
is equivalent tocylc run --icp=POINT SUITE
, andcylc run -w SUITE POINT
is equivalent tocylc run -w --start-point=POINT SUITE
.final-point=POINT
and--stop-point=POINT
option forcylc run
andcylc restart
. The--until=POINT
option is now an alias for--final-point=POINT
option.cylc run
andcylc restart
command line (to override suite configuration.) Previously, it is only possible to disable the setting.To do:
hold point(test already exists)--ignore-final-cycle-point
and the--ignore-stop-cycle-point
options.Requirements check-list
CONTRIBUTING.md
and added my name as a Code Contributor.CHANGES.md
.