----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/53373/#review154557 -----------------------------------------------------------
ambari-server/src/main/java/org/apache/ambari/server/controller/internal/UpgradeResourceProvider.java (lines 1267 - 1273) <https://reviews.apache.org/r/53373/#comment224174> Kind of a weird syntax anyway, can we just use the enhanced iterator of the entrySet? (I know this isn't your code, but while we're here...) ambari-server/src/main/java/org/apache/ambari/server/controller/internal/UpgradeResourceProvider.java (lines 1780 - 1781) <https://reviews.apache.org/r/53373/#comment224177> Wow. Will this always be the correct order? Is our mystery parser making LinkedHashSets or HashSets? ambari-server/src/main/java/org/apache/ambari/server/controller/internal/UpgradeResourceProvider.java (lines 2118 - 2122) <https://reviews.apache.org/r/53373/#comment224173> Could probably skip this and go right to the disjunction check ambari-server/src/main/java/org/apache/ambari/server/controller/internal/UpgradeResourceProvider.java (lines 2138 - 2141) <https://reviews.apache.org/r/53373/#comment224180> Interesting concept. I like the idea, but this will only work if we are one-cluster-at-a-time. If we ever invoke an upgrade against two clusters to the same version and one of them is orchestrating while the other is started (very highly unlikely today) this won't work. I wouldn't be opposed to putting it on the context. Your call. - Nate Cole On Nov. 2, 2016, 8:11 a.m., Jonathan Hurley wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/53373/ > ----------------------------------------------------------- > > (Updated Nov. 2, 2016, 8:11 a.m.) > > > Review request for Ambari, Dmitro Lisnichenko and Nate Cole. > > > Bugs: AMBARI-18685 > https://issues.apache.org/jira/browse/AMBARI-18685 > > > Repository: ambari > > > Description > ------- > > Unit tests are in progress, but seeing as though this is a feature branch, I > didn't think I needed to wait to post the review. I mostly wanted feedback on > the approach before I invested time in the tests. > > Augment the API of creating an Upgrade with the following: > > ``` > POST api/v1/clusters/c1/upgrades > { > "Upgrade": { > "repository_version": "2.5.0.0-965", > "upgrade_type": "HOST_ORDERED", <---------- new value other than > ROLLING/NON_ROLLING > "skip_failures": "false", > "skip_prerequisite_checks": "false", > "skip_manual_verification": "false", > "host_order": [ <------ new > { > "hosts": [ "c6401.ambari.apache.org, "c6402.ambari.apache.org", > "c6403.ambari.apache.org" ], > "service_checks": ["ZOOKEEPER"] > }, > { > "hosts": [ "c6404.ambari.apache.org, "c6405.ambari.apache.org"], > "service_checks": ["ZOOKEEPER", "KAFKA"] > } > ] > } > } > ``` > > {{skip_failures}} must be {{false}} when the upgrade_type is HOST_ORDERED > {{skip_manual_verification}} must be omitted or {{false}} when upgrade_type > is HOST_ORDERED > > > Diffs > ----- > > > ambari-server/src/main/java/org/apache/ambari/server/controller/internal/UpgradeResourceProvider.java > d83aaa2 > > ambari-server/src/main/java/org/apache/ambari/server/state/stack/upgrade/HostOrderGrouping.java > 9c75344 > > ambari-server/src/main/java/org/apache/ambari/server/state/stack/upgrade/HostOrderItem.java > PRE-CREATION > > Diff: https://reviews.apache.org/r/53373/diff/ > > > Testing > ------- > > PENDING > > > Thanks, > > Jonathan Hurley > >
