Abhishek created KUDU-3278:
------------------------------
Summary: DNS entry removal of a tablet server causes one of its
peers to crash
Key: KUDU-3278
URL: https://issues.apache.org/jira/browse/KUDU-3278
Project: Kudu
Issue Type: Improvement
Components: consensus, tserver
Affects Versions: 1.14.0, 1.10.0
Reporter: Abhishek
Steps to replicate:
Let's say a tablet, T1 has three replicas in tablet servers TS1,TS2,TS3.
If TS1 and TS2 are unable to resolve TS3, one of TS1/TS2 ends up crashing
during election/pre-elections irrespective of TS3 state (running/not-running):
Sample failure logs:
W0429 04:14:11.043696 801167 leader_election.cc:270] T
ecf3e9d1608a4d03ac69a09f0df54b9e P b4eb8f7b19dd4b94a313d8674779b350
[CANDIDATE]: Term 9 election: Was unable to construct an RPC proxy to peer
dddc42c5a10b461cb92465815413e996: Network error: unable to resolve address for
achennaka-kudu-4.achennaka-kudu.root.hwx.site: Name or service not known.
Counting it as a 'NO' vote.
F0429 04:14:11.046133 801167 raft_consensus.cc:2743] Check failed: _s.ok() Bad
status: Network error: Could not obtain a remote proxy to the peer.: unable to
resolve address for achennaka-kudu-4.achennaka-kudu.root.hwx.site: Name or
service not known
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
