![]() So far running tests for oslo.db and the SQLAlchemy tests within Nova I spentĪ couple of days figuring out how to get everything to build and run against That shall Not Be Named (except in requirements.txt):Īdded openstack builds to our Jenkins CI system, Within an hour, but we shall now forever see in requirements.txt, the Version Luckily, as I had just learned 30 minutes earlier how to run Nova's tests I wasĪble to drill down to exactly what was going on and SQLAlchemy 0.9.6 was out Really had this effect, only the major jumps saw any issues. ![]() Releases of SQLAlchemy released under OpenStack's lifetime have ever Were failing everywhere and it was due to the 0.9.5 release! Theīizarre coiincidence of this was that none of the thirty-odd point Very unusual not the kind of use I usually do myself but nonetheless SQLAlchemy's 5000 tests but hit a use case within Nova that was not Turns out 0.9.5 had a bad fix in it one that didn't fail any of Was hitting some failures within the FixedIp/FloatingIp tests for which I couldn'tįind the cause, assuming that I just wasn't setting up my environment correctly,Īs Nova's tests are passing for all the Jenkins servers the vast majority of the time.Īs we learned within about 30 minutes of my hitting these problems, it Later that day, as I was learning to run Nova's test suite, I Hundred or so point releases I've put out over nine years, it was a little more overdue Release 0.9.5 of SQLAlchemy was not particularly different than any of the other Out the door, as it had been several months since the last point release, and thenĭo some work learning how Nova's unit tests run and function. On Monday, June 23, I got up in the AM with two goals: get a release of SQLAlchemy ![]() OpenStack presence within SQLAlchemy's CI environment (a.k.a. Not to mention, if you're lookingįor something to do, you can help me implement some of these things! ![]() I should be starting on, that is, which features are most likely to beĪccepted by the community as well as which are the top priority. So that interested parties can chime in with further advice and information,Īnd most importantly to help steer me towards the initial blueprints that My goal here is to share with the community what I've seen and done so far, Is an outline of most of what I've come across and my thoughts going forward. Improvement of exising relational patterns within OpenStackĪpplications, for which oslo.db will serve as the foundation.Īs of June 30, 2014, I've been on the job for about a month, so below Improvement of relational schemas on the OpenStack side, and of course Relational databases aren't suitable to non-relational systems, SQLAlchemy and Alembic directly, migration of projects for which On the table, including new improvements / fixes / features to Within this effort, I'd like to stress that from my POV, everything is Set of practices that are specific to OpenStack's needs. ![]() Improving areas that have been problematic and establishing a very strong SQLAlchemy, Alembic Migrations, and Dogpile caching, to spearhead what isĮssentially an "OpenStack SQLAlchemy 2.0" effort - as I am now full-timeĮmployed by an OpenStack vendor (Red Hat), I've been hired to produce a strongĪnd sustained pass through OpenStack's relational database integration, This page describes the ongoing efforts of Mike Bayer, creator/maintainer of 1.6.8 Transparent Retries of Transactions.1.6.7 Encouraging Bundle / Baked Query Use.1.6.6 Pessimistic Locking - SELECT FOR UPDATE.1.3.1 Eager load and Column load tuning.1.1 OpenStack presence within SQLAlchemy's CI environment (a.k.a. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |