bharathv commented on a change in pull request #1684:
URL: https://github.com/apache/hbase/pull/1684#discussion_r422412820
##########
File path:
hbase-server/src/test/java/org/apache/hadoop/hbase/master/TestMasterShutdown.java
##########
@@ -163,7 +164,16 @@ public void
testMasterShutdownBeforeStartingAnyRegionServer() throws Exception {
assertNotEquals("Timeout waiting for server manager to become
available.",
-1, Waiter.waitFor(htu.getConfiguration(), timeout,
() -> masterThread.getMaster().getServerManager() != null));
- htu.getConnection().getAdmin().shutdown();
+ try {
+ htu.getConnection().getAdmin().shutdown();
Review comment:
Ya, doesn't look like it. This is a known problem I think (Root cause is
a flaky Zk connection. I can repro it locally once every ~100 runs). The
problem here seems to me that shutdown is running inline with the rpc call [1].
Nick actually made the shutdown call async in HBASE-23808 but Stack undid that
in HBASE-24052 [2]. Perhaps the fix is to actually make it async in test until
we fix HBASE-24070? Thoughts? (@saintstack curious what the rationale was in
HBASE-24052).
[1] https://issues.apache.org/jira/browse/HBASE-24070
[2]
https://issues.apache.org/jira/browse/HBASE-24052?focusedCommentId=17068822&page=com.atlassian.jira.plugin.system.issuetabpanels%3Acomment-tabpanel#comment-17068822
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
