Alexey Serbin has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/9755 )

Change subject: [consensus] KUDU-2367 fix replica health report
......................................................................


Patch Set 3:

(6 comments)

http://gerrit.cloudera.org:8080/#/c/9755/3/src/kudu/integration-tests/raft_consensus-itest-base.h
File src/kudu/integration-tests/raft_consensus-itest-base.h:

http://gerrit.cloudera.org:8080/#/c/9755/3/src/kudu/integration-tests/raft_consensus-itest-base.h@85
PS3, Line 85:       const MonoDelta& pre_workload_delay = {});
> I think it would be good to document the new parameters.
Done


http://gerrit.cloudera.org:8080/#/c/9755/3/src/kudu/integration-tests/raft_consensus_nonvoter-itest.cc
File src/kudu/integration-tests/raft_consensus_nonvoter-itest.cc:

http://gerrit.cloudera.org:8080/#/c/9755/3/src/kudu/integration-tests/raft_consensus_nonvoter-itest.cc@1932
PS3, Line 1932: tablet-under-the-test
> nit: maybe clearer to say tablet being tested
Done


http://gerrit.cloudera.org:8080/#/c/9755/3/src/kudu/integration-tests/raft_consensus_nonvoter-itest.cc@1933
PS3, Line 1933: falling
> fall
Done


http://gerrit.cloudera.org:8080/#/c/9755/3/src/kudu/integration-tests/raft_consensus_nonvoter-itest.cc@1937
PS3, Line 1937: // failed replica:
> nit: failed replica is:
Done


http://gerrit.cloudera.org:8080/#/c/9755/3/src/kudu/integration-tests/raft_consensus_nonvoter-itest.cc@1939
PS3, Line 1939: back
> nit: started back up
Done


http://gerrit.cloudera.org:8080/#/c/9755/3/src/kudu/integration-tests/raft_consensus_nonvoter-itest.cc@1994
PS3, Line 1994:     delay = MonoDelta::FromSeconds(3 * kUnavaiableFailedSec);
> not quite sure i understand the purpose of this delay
That's to make the leader replica to report the state of the tablet as FAILED.  
Later on, when the replica falls behind the WAL segment GC threshold, the 
leader replica should report the status as FAILED_UNRECOVERABLE.

The bug in the code of the PeerMessageQueue::PeerHealthStatus() method resulted 
in having the leader reporting the health status of the replica still as FAILED 
even after the replica fell behind the WAL segment GC threshold.

I'll add a comment.



--
To view, visit http://gerrit.cloudera.org:8080/9755
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: kudu
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: If77474adc8f618f2cda35f992190133138fdb511
Gerrit-Change-Number: 9755
Gerrit-PatchSet: 3
Gerrit-Owner: Alexey Serbin <[email protected]>
Gerrit-Reviewer: Alexey Serbin <[email protected]>
Gerrit-Reviewer: Jean-Daniel Cryans <[email protected]>
Gerrit-Reviewer: Kudu Jenkins
Gerrit-Reviewer: Mike Percy <[email protected]>
Gerrit-Reviewer: Tidy Bot
Gerrit-Reviewer: Todd Lipcon <[email protected]>
Gerrit-Comment-Date: Fri, 23 Mar 2018 01:22:14 +0000
Gerrit-HasComments: Yes

Reply via email to