[JUJU-523] Ensure a robust isController wrapper for machine manifolds #13668
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
ifDatabaseUpgradeComplete - it is already implied, and therefore redundant.
manadart
changed the title
jameinel
reviewed
Jan 28, 2022
watcher instead of state. This means that when a state ping fails, controller-gated workers that do not depend on state will not bounce.
manadart
force-pushed
the
2.9-robust-iscontroller-flag
branch
from
da7b618
to
8763353
Compare
manadart
force-pushed
the
2.9-robust-iscontroller-flag
branch
from
8763353
to
9f5bdf9
Compare
jameinel
approved these changes
Jan 31, 2022
|
|
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Some manifolds in the machine dependency engine are wrapped with
isControllerto ensure they only run on controllers. This wrapper in turn has the state worker as a dependency.Depending on state in this case is problematic, because if the state fails to ping Mongo, all such workers will bounce whether or not they depend on state.
Instead we use the state config watcher, which determines whether we are a controller based on the presence of state serving info. This will remain static when we change the Mongo primary (and bounce the state worker), leaving fewer workers to restart in such an event.
In addition, some redundant usage of
isControlleris removed where it is already implied by other dependencies.QA steps
juju enable-ha.juju deploy ubuntu.juju ssh 0and runjuju_engine_report.juju ssh -m controller 0and runjuju_engine_report.rs.stepDown()In my simple A/B test, these workers did not bounce for a Mongo primary switch:
Documentation changes
None.
Bug reference
N/A