> On March 9, 2015, 5:50 p.m., Maxim Khutornenko wrote: > > src/main/java/org/apache/aurora/scheduler/updater/JobUpdateControllerImpl.java, > > line 291 > > <https://reviews.apache.org/r/31814/diff/2/?file=888120#file888120line291> > > > > What's the motivation behind dropping the "Aurora Updater" here? > > Bill Farner wrote: > I think the magic user value abuses a field that serves a different > purpose. IMHO an API consumer should be able to programmatically determine > that the scheduler independently performed an action without resorting to > string matching on the magic value of a field. > > In this particular case, i don't think the user field adds signal to what > is already present in the state enum value. > > Maxim Khutornenko wrote: > I still think having a special "Aurora Updater" user clearly visible in > the UI makes it way easier to grasp the event origin rather than trying to > decipher a particular state meaning. Users may not and should not be familar > with the update state diagram in order to understand the cause of the > transition. This is especially true in a multi-actor environment where state > transitions may come from users (pause/resume/abort), external monitoring > service (pause) or the updater itself (pulse expired). > > Bill Farner wrote: > > I still think having a special "Aurora Updater" user clearly visible in > the UI > > I completely agree, but the UI should make that decision, not the > scheduler. > > Maxim Khutornenko wrote: > I see, sure I am fine with that. Do you propose to address it separately? > If so, mind filing a ticket? > > Kevin Sweeney wrote: > +1 to removal of the magic string - hacking the ui should be done in the > ui.
Went with a more broad ticket to cover this: https://issues.apache.org/jira/browse/AURORA-1178 - Bill ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/31814/#review75726 ----------------------------------------------------------- On March 7, 2015, 1:06 a.m., Bill Farner wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/31814/ > ----------------------------------------------------------- > > (Updated March 7, 2015, 1:06 a.m.) > > > Review request for Aurora, Joshua Cohen and Maxim Khutornenko. > > > Bugs: AURORA-1077 > https://issues.apache.org/jira/browse/AURORA-1077 > > > Repository: aurora > > > Description > ------- > > Include messages with internal job updater state transitions. > > > Diffs > ----- > > src/main/java/org/apache/aurora/scheduler/updater/InstanceUpdater.java > 09f2a74c9e6f6dbf79f83aa7a387587c593b9f0e > > src/main/java/org/apache/aurora/scheduler/updater/JobUpdateControllerImpl.java > acdade3dca807a221b4da975d0310c91884ee752 > src/main/java/org/apache/aurora/scheduler/updater/OneWayJobUpdater.java > 27a5b9026f5ac3b3bdeb32813b10435bc3dab173 > src/main/java/org/apache/aurora/scheduler/updater/SideEffect.java > 27e0654bfb90f48b407edda5a0c914e595d9c552 > src/main/java/org/apache/aurora/scheduler/updater/StateEvaluator.java > 66b0e4b7a9b23e3e6c772a21f5adc39e1d1461ad > src/test/java/org/apache/aurora/scheduler/updater/InstanceUpdaterTest.java > 4db0080547d61af1511a4fb62bf88b3bbf819f1e > src/test/java/org/apache/aurora/scheduler/updater/JobUpdaterIT.java > e24d6bde5f3479a75522e825cce4ec6c30c117aa > src/test/java/org/apache/aurora/scheduler/updater/OneWayJobUpdaterTest.java > 7d0a7438b4a517e5e0d44f4e99aceb1a6d19f987 > > Diff: https://reviews.apache.org/r/31814/diff/ > > > Testing > ------- > > > Thanks, > > Bill Farner > >