September 07, 2022
Overview
- Katello Schedule
Pulp
- Pulpcore
- Query of all changes from last week
- 3.21.0 will release tomorrow
- “The” deadlock fix backported for curr-Y Satellite
- RPM
- 3.18 blockers in the process of being merged
- Still minor cleanup to finish up pre-release
- Considering some changes that would allow pulp_rpm 3.18 to be used alongside pulpcore 3.18
- Currently, as mentioned last week, it contains a feature that relies on pulpcore 3.19
- Coordination going on with certain BZs / customer cases, hotfixes
- 3.18 blockers in the process of being merged
- Migration
- Ansible
- Pulp Container
- OSTree
- Pulp CLI
Katello
- Testathon in-progress, so changes are on-hold for two weeks
- Waiting patiently for core/3.21
- RDU Tean meetups next week - should coord w/ pulp RDU meetup
QE
- next-Z snap 1 bz verification underway
- next-Y snap 9 bz verification
- Feature testing
- Automation failure fixes
- Waiting on static ouia-id to get into a build
- Some missing OUIA-ID for cv, sent BZ back to dev
- Pending - Host details page OUIA-ID PR
- Question: how long is “too long” to publish a CV where we have added 1 erratum and incrementally updating (ie, we’re doing depsolving)
- Rhel8_baseos + Rhel8_appstream
- Created a cv with a filter of before June 30th
- Publish and did incremental update
- Incremental update time: 10min 15 sec
- Publishing cv after inc update: 10min 15 sec
- Created a filter that includes up to July 31st
- Publish and did incremental update
- Incremental update time: 10min 15 sec
- Publishing cv after inc update: 10min 15 sec
- rhel7
- Created a cv with a filter of before June 30th
- Publish and did incremental update
- Incremental update time: 9min 9 sec
- Publishing cv after inc update: 13min 23 sec
- Created a filter that includes up to July 31st
- Publish and did incremental update
- Incremental update time: 8min 49 sec
- Publishing cv after inc update: 18min 35 sec
- [Iballou] It would be interesting to see the breakdown between “pulp time” and “katello time” for this
- Katello-indexing should be much faster now, would be good to know if that isn’t the case
- Note: depSolving is not selected for these runs, but iirc, doing filtering and publishing does the depsolving?
- In general usage, it does not - katello only asks Pulp to depsolve, if the user asks for it explicitly
- Incrementals depsolve by default, so they do
- Rhel8_baseos + Rhel8_appstream