Skip to content
This repository was archived by the owner on May 25, 2026. It is now read-only.

rev101

  • rev101
  • 13c1ad5
  • Verified

    This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
  • Choose a tag to compare

  • rev101
  • 13c1ad5
  • Choose a tag to compare

  • Verified

    This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
@paulomach paulomach tagged this 07 Feb 14:17
* workflow updates

* libs updates

* continuous write test charm

* replicating function and name from k8s

* downstream updates

* refactor replication tests with write load

* line length fix

* consistent test structure across platforms

* moving tests structure

* cont. writes fixture

* add new method

* refactor update_endpoints adding periodic checks to handle non event related changes

* lintfixes

* refactor first tests to add load

* wont block for transient states

* add continuos writes to remaining tests

* test fixes

* ensure using 1st endpoint

* sst test now purges binlogs

* flush and purge

* increase time generosity

* retry to accommodate daemon initialization

* CI adaptation

* avoid getting trapped by unresponsive call

* speed up loop

* increased verbosity

* std timeouts

* missing commits

* temporary CI access to debug test run

* update lib, import error related to it

* fixes for reboot support

* bump crypto to fix import error

* remove tmate ci session

* bit faster attempts

* workaround for lxd containers failing to mount storage

* fix typos

* happy linting

* avoid unintend host reboot

* log failed storage as error

* add missing `series` and trying faster pack

* refactoring and formatting

* fix for tox v4

* allow insert value to be incremented - avoid PK conflict

* treat exception for teardown case

* avoid on_start run on on_install problem

* flakiness mitigations

* more flakiness stripping

* docstring

* `beta` channel for tls

* further hardening

* pack one series at the time

* allow more complexity

* more robust way to wait recover

* more time

* even more time

* yet even more time

* moving complexity out of on_start

* addressing pr comments

* never force reboot the cluster

* defer runs of leader elected

* skip teardown for non initilized unit

* fix typo

* wait unit full initialization

* trigger update status after reboot

* fix typo

* fix peer data reference

* linting fixes

* edge cases treatment and unit tls flag to avoid restarts

* set control flag to enable cert/key rotation

* pr comments

* remove ambiguity by validating event_unit_data
Assets 2
Loading