-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/37739/
-----------------------------------------------------------
Review request for Ambari, Alejandro Fernandez and Nate Cole.
Bugs: AMBARI-12867
https://issues.apache.org/jira/browse/AMBARI-12867
Repository: ambari
Description
-------
On 1000 node RU I had 2.3.0.0-2557 installed with some 20 hosts down with
heartbeat lost. Then I registered 2.3.2.0-2664 and when I proceeded to install,
it would always get aborted with no logs in server or agents.
Turns out that whenever we install, we do so in stages containing 100 hosts
each. If any of the host failed or timed out etc., the rest of the stages are
aborted. So in this case the first stage had 1 host timeout, which resulted in
that and other stages being aborted.
I cannot install a version without all hosts being alive. Workaround seems to
be to delete lost hosts from Ambari.
The solution is to use the stage's success criteria to determine if the other
stages in the request should be aborted.
Diffs
-----
ambari-server/src/main/java/org/apache/ambari/server/Role.java 636df3f
ambari-server/src/main/java/org/apache/ambari/server/controller/internal/ClusterStackVersionResourceProvider.java
6133885
ambari-server/src/test/java/org/apache/ambari/server/controller/internal/ClusterStackVersionResourceProviderTest.java
a56823b
Diff: https://reviews.apache.org/r/37739/diff/
Testing
-------
mvn clean test
Thanks,
Jonathan Hurley
