Use juju-run to invoke juju-reboot for unit reboot #575
Conversation
Issue #921 has the context for the change: in order to avoid triggering reboots asynchronously to juju hook executions (when `juju ssh reboot` is done). openstack-charmers/zaza-openstack-tests#921
| generic_utils.juju_reboot(_unit) | ||
| self.subprocess.check_call.assert_called_once_with( | ||
| ['juju', 'ssh', _unit, | ||
| f'sudo juju-run -u {_unit} "juju-reboot --now"']) |
There was a problem hiding this comment.
We tend not to use f-strings in zaza purely due to the python3.5 support. We need to formally make a break on it, but zaza still tentatively supports py35!
There was a problem hiding this comment.
Addressed in 7417002
There was a problem hiding this comment.
This test code still uses the f-string? I see the code-under-test does use .format().
|
Now all the functional tests are failing, which is unrelated to the changes. |
Codecov ReportBase: 89.22% // Head: 89.19% // Decreases project coverage by
Additional details and impacted files@@ Coverage Diff @@
## master #575 +/- ##
==========================================
- Coverage 89.22% 89.19% -0.03%
==========================================
Files 44 44
Lines 4594 4610 +16
==========================================
+ Hits 4099 4112 +13
- Misses 495 498 +3
Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here. ☔ View full report at Codecov. |
|
I gave it a try here https://review.opendev.org/c/x/charm-ovn-dedicated-chassis/+/857878/2#message-37afbdf3f745d77719aa43afe5234aaa87d70468 but got an unrelated failure which actually has to do with a regression I introduced to charm-layer-ovn (openstack-charmers/charm-layer-ovn#71 - a fix). When a fix is landed I'll give it another go. |
Okay, sounds good. |
|
Raised https://review.opendev.org/c/x/charm-ovn-chassis/+/858491 to help test this change instead of the in-flight review on the ovn-dedicated-chassis charm. |
fnordahl
dismissed
ajkavanagh’s stale review
To me it appears the author has addressed the review comments falling back to .format in the runtime code, and I know you're busy with other stuff. So in the interest of foreward movement I'll push forward. You're right in the unit test containing a f-string, and the gate check for zaza itself is Python 3.6+.
Issue #921 has the context for the change: in order to avoid triggering
reboots asynchronously to juju hook executions (when
juju ssh rebootis done).
openstack-charmers/zaza-openstack-tests#921