Sprint taskboard:
Here is a summary of what the team did from Jan 25 to Feb 8.

Regarding Software Factory:
We have played around with Kerberos, leading to the grooming of a
future implementation:
We have move the resources engine API under the new managesf API v2 endpoint
We have investigated a solution to make the resources engine apply
call asynchronuous and discuss with team and finally decided a new
solution w/o passing through the manageSF API
We have changed the sf-ops backup playbook to facilitate old backup
deletion with bup.
We have boostrapped a website for Software Factory with static pages
and blog based on pelican
We investigated gerrit-2.14 integration, though database schema
changed and managesf/initialization needs to be adapted to use the
REST api
We integrated ara in sfconfig so that all ansible-playbook execution are stored
We removed zuulV2/nodepoolV2 and Jenkins and dropped the version
prefix in zuul/nodepool.
We updated the zuul and nodepool package to use the new Javascript
tooling stack.
We added dynamic collections of zuul and nodepool configuration to
generate grafana dashboard.
We abandoned the graphite integration to avoid to have two metrics
backends in SF. We started to prepare a script to automatically
generate nodepool and zuul graph based on yaml file.

Regarding RepoXplorer:

Regarding openstack-infra:
We worked on upstream patches:
    - nodepool max-hold-age setting https://review.openstack.org/#/c/536295/
    - nodepool admin webapp https://review.openstack.org/#/c/536319/
    - zuul autohold filtering per ref/commit (abandoned in favor of an
earlier patch) https://review.openstack.org/#/c/540035/
    - zuul admin webapp https://review.openstack.org/#/c/539004/
We got the nodepool static provider merged as well as the drivers
plugin interface
We worked on Zookeeper command retry integration:
We proposed a couple of Summit talk: SF as third-party-ci and Anomaly
detection workshop
There was a IPv6 routing problem on some VMs upstream which ended up
requiring Rackspace's attention to fix
There was a Zookeeper disconnection which snowballed into different
issues: https://etherpad.openstack.org/p/HRUjBTyabM
CityCloud has asked us to stop using their for nodepool capacity until
July (-100 VMs)

Regarding other things:
We triaged a couple of VMT bug
We presented SF to the University of Quebec in Montreal (UQAM). They
want to integrate a CI/CD workflow, but don't use ansible, git or
gerrit, but they are interessted with the workflow. Also they use

The meeting can be watched at https://bluejeans.com/s/Eqgam
Next public review will happen Thursday Feb 22 13:00 UTC, feel free to
join us on #softwarefactory for the event details.

The SF team

Softwarefactory mailing list

Reply via email to