cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Thibaut (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-2543) Node not responding, bringing down cluster, marked as up
Date Fri, 22 Apr 2011 14:46:06 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-2543?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13023232#comment-13023232
] 

Thibaut commented on CASSANDRA-2543:
------------------------------------

Ok thanks.

I will try above to make sure it's the GC thread taking up 100% cpu (which I suspect).

The GC indeed ran indeed very very often: (grepping the old log file)

 INFO [ScheduledTasks:1] 2011-04-22 15:07:24,742 GCInspector.java (line 128) GC for ConcurrentMarkSweep:
3133 ms, 1537221688 reclaimed leaving 3830570400 used; max is 5605687296
 INFO [ScheduledTasks:1] 2011-04-22 15:07:28,978 GCInspector.java (line 128) GC for ConcurrentMarkSweep:
2901 ms, 2172093472 reclaimed leaving 3219903248 used; max is 5605687296
 INFO [ScheduledTasks:1] 2011-04-22 15:07:33,946 GCInspector.java (line 128) GC for ConcurrentMarkSweep:
3138 ms, 1715631768 reclaimed leaving 3581355456 used; max is 5605687296
 INFO [ScheduledTasks:1] 2011-04-22 15:07:38,318 GCInspector.java (line 128) GC for ConcurrentMarkSweep:
2860 ms, 2467501736 reclaimed leaving 2874209912 used; max is 5605687296
 INFO [ScheduledTasks:1] 2011-04-22 15:07:43,466 GCInspector.java (line 128) GC for ConcurrentMarkSweep:
3059 ms, 1491300856 reclaimed leaving 3875493760 used; max is 5605687296
 INFO [ScheduledTasks:1] 2011-04-22 15:07:47,529 GCInspector.java (line 128) GC for ConcurrentMarkSweep:
2780 ms, 2521501408 reclaimed leaving 2824239464 used; max is 5605687296
 INFO [ScheduledTasks:1] 2011-04-22 15:07:52,983 GCInspector.java (line 128) GC for ConcurrentMarkSweep:
3330 ms, 1287324312 reclaimed leaving 3964465048 used; max is 5605687296
 INFO [ScheduledTasks:1] 2011-04-22 15:07:57,147 GCInspector.java (line 128) GC for ConcurrentMarkSweep:
3071 ms, 1950115232 reclaimed leaving 3337420088 used; max is 5605687296
 INFO [ScheduledTasks:1] 2011-04-22 15:08:01,785 GCInspector.java (line 128) GC for ConcurrentMarkSweep:
3069 ms, 1748108592 reclaimed leaving 3444685288 used; max is 5605687296
 INFO [ScheduledTasks:1] 2011-04-22 15:08:06,397 GCInspector.java (line 128) GC for ConcurrentMarkSweep:
3011 ms, 1779092920 reclaimed leaving 3637964728 used; max is 5605687296
 INFO [ScheduledTasks:1] 2011-04-22 15:08:10,715 GCInspector.java (line 128) GC for ConcurrentMarkSweep:
2950 ms, 2042754088 reclaimed leaving 3223819496 used; max is 5605687296
 INFO [ScheduledTasks:1] 2011-04-22 15:08:15,528 GCInspector.java (line 128) GC for ConcurrentMarkSweep:
3217 ms, 956021816 reclaimed leaving 4371838768 used; max is 5605687296
 INFO [ScheduledTasks:1] 2011-04-22 15:08:19,380 GCInspector.java (line 128) GC for ConcurrentMarkSweep:
3007 ms, 2216336016 reclaimed leaving 3213552904 used; max is 5605687296
 INFO [ScheduledTasks:1] 2011-04-22 15:08:24,154 GCInspector.java (line 128) GC for ConcurrentMarkSweep:
2973 ms, 1964499424 reclaimed leaving 3287557528 used; max is 5605687296
 INFO [ScheduledTasks:1] 2011-04-22 15:08:29,088 GCInspector.java (line 128) GC for ConcurrentMarkSweep:
2988 ms, 1964657952 reclaimed leaving 3418170312 used; max is 5605687296


It still interest me why the node didn't recover even though there were nearly no requests
at all from our application.



> Node not responding, bringing down cluster, marked as up
> --------------------------------------------------------
>
>                 Key: CASSANDRA-2543
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-2543
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Core
>    Affects Versions: 0.7.4
>            Reporter: Thibaut
>             Fix For: 0.7.6
>
>         Attachments: jstack
>
>
> I have one node which constantly hangs and brings done the entire cluster (not giving
any answers).
> If I restart the node, the node will hang after a certain number of time. I have no indication
> It's marked as up when executing the nodetool ring command.
> Executing the ring command on the node itself (without any traffic on the cluster) takes
at least 2 minutes to execute. The node takes about 50%-100% of cpu over all cpus.
> Netstats doesn't list anything interesting:
> /software/cassandra/bin/nodetool -h localhost netstats
> Mode: Normal
> Not sending any streams.
> Not receiving any streams.
> Pool Name                    Active   Pending      Completed
> Commands                        n/a         0          51064
> Responses                       n/a         0         530479
> I attached the jstack of the node. There are no indications that the node has faulty
hardware. 
> /usr/bin/java -ea -XX:+UseThreadPriorities -XX:ThreadPriorityPolicy=42 -Xms5254M -Xmx5254M
-Xmn400M -XX:+HeapDumpOnOutOfMemoryError -Xss128k -XX:+UseParNewGC -XX:+UseConcMarkSweepGC
-XX:+CMSParallelRemarkEnabled -XX:SurvivorRatio=8 -XX:MaxTenuringThreshold=1 -XX:CMSInitiatingOccupancyFraction=75
-XX:+UseCMSInitiatingOccupancyOnly -Djava.net.preferIPv4Stack=true -Dcom.sun.management.jmxremote.port=8080
-Dcom.sun.management.jmxremote.ssl=false -Dcom.sun.management.jmxremote.authenticate=false
-Dlog4j.configuration=log4j-server.properties -Dlog4j.defaultInitOverride=true -Dcassandra-foreground=yes
-cp /software/cassandra/bin/../conf:/software/cassandra/bin/../build/classes:/software/cassandra/bin/../lib/antlr-3.1.3.jar:/software/cassandra/bin/../lib/apache-cassandra-0.7.4.jar:/software/cassandra/bin/../lib/avro-1.4.0-fixes.jar:/software/cassandra/bin/../lib/avro-1.4.0-sources-fixes.jar:/software/cassandra/bin/../lib/commons-cli-1.1.jar:/software/cassandra/bin/../lib/commons-codec-1.2.jar:/software/cassandra/bin/../lib/commons-collections-3.2.1.jar:/software/cassandra/bin/../lib/commons-lang-2.4.jar:/software/cassandra/bin/../lib/concurrentlinkedhashmap-lru-1.1.jar:/software/cassandra/bin/../lib/guava-r05.jar:/software/cassandra/bin/../lib/high-scale-lib.jar:/software/cassandra/bin/../lib/jackson-core-asl-1.4.0.jar:/software/cassandra/bin/../lib/jackson-mapper-asl-1.4.0.jar:/software/cassandra/bin/../lib/jetty-6.1.21.jar:/software/cassandra/bin/../lib/jetty-util-6.1.21.jar:/software/cassandra/bin/../lib/jline-0.9.94.jar:/software/cassandra/bin/../lib/json-simple-1.1.jar:/software/cassandra/bin/../lib/jug-2.0.0.jar:/software/cassandra/bin/../lib/libthrift-0.5.jar:/software/cassandra/bin/../lib/log4j-1.2.16.jar:/software/cassandra/bin/../lib/servlet-api-2.5-20081211.jar:/software/cassandra/bin/../lib/slf4j-api-1.6.1.jar:/software/cassandra/bin/../lib/slf4j-log4j12-1.6.1.jar:/software/cassandra/bin/../lib/snakeyaml-1.6.jar
org.apache.cassandra.thrift.CassandraDaemon

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message