[
https://issues.apache.org/jira/browse/RATIS-2487?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Tsz-wo Sze updated RATIS-2487:
------------------------------
Attachment: 1420_review.patch
> NPE when leader previous log is already purged
> ----------------------------------------------
>
> Key: RATIS-2487
> URL: https://issues.apache.org/jira/browse/RATIS-2487
> Project: Ratis
> Issue Type: Bug
> Reporter: Ivan Andika
> Assignee: Ivan Andika
> Priority: Major
> Attachments: 1420_review.patch, ratis-2487-npe.txt
>
> Time Spent: 20m
> Remaining Estimate: 0h
>
> Found the following logs in test where one thread restart different leaders
> multiple times and another thread kept sending writes.
> {code:java}
> java.lang.NullPointerException: omNode-2@group-523986131536->omNode-1:
> Previous TermIndex not found for firstIndex = 93 at
> java.base/java.util.Objects.requireNonNull(Objects.java:360) at
> org.apache.ratis.server.leader.LogAppenderBase.assertProtos(LogAppenderBase.java:270)
> at
> org.apache.ratis.server.leader.LogAppenderBase.newAppendEntriesRequest(LogAppenderBase.java:255)
> at
> org.apache.ratis.grpc.server.GrpcLogAppender.appendLog(GrpcLogAppender.java:387)
> at
> org.apache.ratis.grpc.server.GrpcLogAppender.run(GrpcLogAppender.java:262)
> at
> org.apache.ratis.server.leader.LogAppenderDaemon.run(LogAppenderDaemon.java:80)
> at java.base/java.lang.Thread.run(Thread.java:1583) {code}
> Seems to be related to this Ratis NPE when there is no previous log (when the
> logs are purged) when raft.server.log.purge.upto.snapshot.index is true.
> After RATIS-2427, NPE will not cause LogAppender to be restarted. This can
> cause the leader to not send anything anymore and can cause the Raft group to
> be stuck.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
