Last Week in ManageIQ: Swift, vCloud, PG upgrades, and the Euwe beta
Hi everyone! This is Jason, bringing you this week’s Last Week in ManageIQ. This past week was the final development week before the upcoming beta release of the Euwe branch! After the beta release, we will start on many weeks of stabilization, so keep on the lookout for lots of bug fixes in the coming weeks.
Last week, we merged 106 PRs, of which 37 were enhancements. In total, there were 339 commits by 47 contributors. If you want to join in on the fun, be sure to checkout the issues tracker.
Move the postgresql mountpoint up
Along with ManageIQ/manageiq-appliance-build#165
this PR changes the mountpoint location of the PostgreSQL data disk. The
pg_upgrade tool from PostgreSQL has a really cool feature where it will use
hard links for files that are the same on both sides of the upgrade, which can
save a significant amount of space. However, the current mountpoint is too deep
and makes taking advantage of that feature difficult. So, this PR moves us in a
direction for easier PostgreSQL upgrades. Thanks
Handle the case that a parameter group does not have label
OpenStack Heat templates are supposed to follow the spec, and part of the spec is to include a label and description for parameter groups. However, in reality, not every template has these properties. Bill put together this PR to loosen up the requirements, making it just that much easier for our users.
Disable same origin as host matching in development
Dávid wrote up this pull request for one of those is-it-a-fix-or-an-enhancement bugs, where he loosened the same-origin policy in development mode only. Developers who work on the service UI will benefit the most, as they are once again able to run the service UI in development mode and still have it talk over websockets on another host. In production, on the appliance, the same-origin policy still applies and has not been changed.
Force expand node when selecting service catalog entry point
Coincidentally, Dávid has the second featured fix this week, with a fix to the service catalog entry point modal. The change is to automatically expand the tree to the point of the current selection. Sometimes a one-line fix goes a long way towards usability!
New Swift Storage Provider
As we expand further into managing storage providers, Jerry has added management of OpenStack Swift object storage. Along with the recent Cinder addition, this greatly improves ManageIQ’s storage coverage. Moving forward, this Swift integration could be used as the base for non-OpenStack standalone Swift providers.
Implement VMware Cloud EventCatcher backend
Miha expanded on the VMware vCloud provider with this enhancement to add the EventCatcher. When creating a new provider, the EventCatcher is one of the most crucial components, as it is generally the main driver of automatic inventory updates, so this is a really important update for vCloud support!
Extract ActsAsArModel to QueryRelation gem
Although not technically a true “deletion”, the ActsAsArQuery functionality was extracted to the query_relation gem by yours truly. ActsAsArQuery, originally written by Keenan, provides an ActiveRecord-like query interface for non-SQL backends. Moving it out to a new gem allows it to be reused by the new manageiq-api-client.
Drop redundant User.in_my_region
When enterprise regioning was in its early days, the User model had its own region column. As time went on, all tables assumed their region number from the id range and general methods were made for querying by region. However, the User model never changed and has been that way since 2010! Šimon finally got rid of these in this pull request, and has made the User model consistent with everything else.
It was really hard for me to choose this week, but take a look here for all of the other merged PRs. Have a great week and Happy Coding!