[
https://issues.apache.org/jira/browse/HADOOP-4051?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12627148#action_12627148
]
Koji Noguchi commented on HADOOP-4051:
--------------------------------------
When taking jstack of 0.17 namenode, it showed many handlers with
{noformat}
"IPC Server handler 38 on 8020" daemon prio=10 tid=0x00002aaf84013000
nid=0x403b waiting for monitor entry [0x0000000044b52000..0x0000000044b52b00]
java.lang.Thread.State: BLOCKED (on object monitor)
at
org.apache.hadoop.dfs.FSNamesystem.blockReportProcessed(FSNamesystem.java:1967)
- waiting to lock <0x00002aab7184f418> (a
org.apache.hadoop.dfs.FSNamesystem)
at org.apache.hadoop.dfs.NameNode.sendHeartbeat(NameNode.java:571)
at sun.reflect.GeneratedMethodAccessor12.invoke(Unknown Source)
at
sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)
at java.lang.reflect.Method.invoke(Method.java:597)
at org.apache.hadoop.ipc.RPC$Server.call(RPC.java:446)
at org.apache.hadoop.ipc.Server$Handler.run(Server.java:896)
{noformat}
> Impact in NameNode scalability because heartbeat processing acquires the
> global lock
> ------------------------------------------------------------------------------------
>
> Key: HADOOP-4051
> URL: https://issues.apache.org/jira/browse/HADOOP-4051
> Project: Hadoop Core
> Issue Type: Bug
> Components: dfs
> Affects Versions: 0.17.0, 0.17.1, 0.17.2, 0.18.0
> Reporter: dhruba borthakur
>
> The heartbeat processing code recently got rearranged via HADOOP-3254. This
> caused the NameNode heartbeat processing code to acquire the FSNamesystem
> global lock for every heartbeat processing. This could impact scalability of
> the namenode.
> This problem is present in 0.17.x and 0.18.x release only. It is not present
> in 0.16.x and on trunk.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
