[HDDS-1555] Disable install snapshot for ContainerStateMachine

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.3.0
• Fix Version/s: 0.5.0
• Component/s: Ozone Datanode
• Labels: MiniOzoneChaosCluster pull-request-available
• Sprint: Done
• Assignee: Siddharth Wagle
• Reporter: Mukul Kumar Singh
• Created: 19/May/19 05:32
• Updated: 24/Feb/20 20:52
• Resolved: 01/Jul/19 15:30

Description

In case a follower lags behind the leader by a large number, the leader tries to send the snapshot to the follower. For ContainerStateMachine, the information in the snapshot it not the entire state machine data. InstallSnapshot for ContainerStateMachine should be disabled.

2019-05-19 10:58:22,198 WARN  server.GrpcLogAppender (GrpcLogAppender.java:installSnapshot(423)) - GrpcLogAppender(e3e19760-1340-4acd-b50d-f8a796a97254->28d9bd2f-3fe2-4a69-8120-757a00fa2f20): failed to install snapshot [/Users/msingh/code/apache/ozone/github/git_oz_bugs_fixes/hadoop-ozone/integration-test/target/test/data/MiniOzoneClusterImpl-c2a863ef-8be9-445c-886f-57cad3a7b12e/datanode-6/data/ratis/fb88b749-3e75-4381-8973-6e0cb4904c7e/sm/snapshot.2_190]: {}
java.lang.NullPointerException
        at org.apache.ratis.server.impl.LogAppender.readFileChunk(LogAppender.java:369)
        at org.apache.ratis.server.impl.LogAppender.access$1100(LogAppender.java:54)
        at org.apache.ratis.server.impl.LogAppender$SnapshotRequestIter$1.next(LogAppender.java:318)
        at org.apache.ratis.server.impl.LogAppender$SnapshotRequestIter$1.next(LogAppender.java:303)
        at org.apache.ratis.grpc.server.GrpcLogAppender.installSnapshot(GrpcLogAppender.java:412)
        at org.apache.ratis.grpc.server.GrpcLogAppender.runAppenderImpl(GrpcLogAppender.java:101)
        at org.apache.ratis.server.impl.LogAppender$AppenderDaemon.run(LogAppender.java:80)
        at java.lang.Thread.run(Thread.java:748)

Issue Links

• duplicates: HDDS-1441 Remove usage of getRetryFailureException
• is blocked by: RATIS-592 One node ratis writes fail forever after first NotLeaderException or LeaderNotReadyException
• links to: GitHub Pull Request #846