Hi everyone, below you can find the updates from the Canonical Juju Team for the week starting on the 18/07/2022. Have a great weekend!
- Reduced charmhub package’s exported API surface and simplified how new Charmhub clients are created
- Suggested improvements to the Juju project PR description template
- Update deploy_revision tests to actually check revision
- Fix cli/local_charms test due to deleted GitHub repo
- Make charmhub download test use postgresql as mysql doesn’t have stable
- Pebble: reviewed Robert’s draft log forwarding PR
- Pebble: discussions on issues related to exec environment variables: #125, #126, #127
- mid cycle planning sprint
- fix timing issue removing a unit with a lxd profile
- fix timing issue deploying a charm with an lxd profile
- fix upgrade info for subordinate units
add new option for
- fix issue bootstrapping on azure with custom resource group
remove more uses of
I have been working on a PR for moving
upgrade-controller/model logic to the controller side and also rebuilt the sync-tool command - https://github.com/juju/juju/pull/14344
This week I spent my time looking at charms used for testing within Juju. This involved moving these charms over to Charmhub support and centralizing their location and updating tests as needed. Work is ongoing and PR incoming.
- Fixed a bug in 3.0-beta2 where
juju refreshextremely quickly after
juju deploycould incorrectly return no change needed. Seen during test.
- Fixed: 3.0-beta2 deploy with --resource-rev fails: ERROR empty charm ID
- Fixed a bug where the actual resource downloaded by juju was not the requested one at deploy. Follow on to bug above.
- Fixed 3.0-beta2 charms not downloaded when deployed with --force for series
- juju 3.0 beta3 release
- fix for removing a machine from a HA controller cluster
- Patch that prevents excessive calls to the provider, attempting to retrieve network configuration when no instances are found.
- Follow up to the patch above that activates a back-off for machines with unrecognised instances.
- Enery consumption enhancement to ensure that we don’t activate LXD unless we are provisioning containers on a machine.