Hello Kudu Jenkins,
I'd like you to reexamine a change. Please visit
http://gerrit.cloudera.org:8080/17691
to look at the new patch set (#2).
Change subject: [master] make automatic AddMaster resilient to leadership
changes
......................................................................
[master] make automatic AddMaster resilient to leadership changes
AutoAddMasterTest.TestAddWithOnGoingDdl would previously fail with the
new master shutting down with the following error:
Remote error: Failed to perform AddMaster RPC: Illegal state: Failed initiating
master Raft ChangeConfig request, error: unknown: Leader has not yet committed
an operation in its own term
This patch addresses this by catching errors commonly seen in the Raft
layer and retrying, as we might do for DML operations.
Without this patch, the test failed ~10% of the time in debug mode with
--stress_cpu_threads=3. With the patch, the test only fails ~1% of the
time, which will be addressed in a later patch.
This patch also addresses some leftover feedback from
7e66534d0e62fb850bf300d52da4a0a76889f4b8 regarding verification of
masters in the presence of DNS failures. I didn't add a new test for
this because it turned out to not change the behavior of things -- the
new master would still fail upon attempting to resolve replicas.
Change-Id: Ie38453c6fc41ce98c59c010902e2d9fe9db62dee
---
M src/kudu/master/dynamic_multi_master-test.cc
M src/kudu/master/master.cc
M src/kudu/master/master_runner.cc
M src/kudu/master/master_service.cc
4 files changed, 38 insertions(+), 12 deletions(-)
git pull ssh://gerrit.cloudera.org:29418/kudu refs/changes/91/17691/2
--
To view, visit http://gerrit.cloudera.org:8080/17691
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings
Gerrit-Project: kudu
Gerrit-Branch: master
Gerrit-MessageType: newpatchset
Gerrit-Change-Id: Ie38453c6fc41ce98c59c010902e2d9fe9db62dee
Gerrit-Change-Number: 17691
Gerrit-PatchSet: 2
Gerrit-Owner: Andrew Wong <[email protected]>
Gerrit-Reviewer: Kudu Jenkins (120)
