Can't run Solr wikipedia sample in cluster

(5 posts) (3 voices)

• Started 8 months ago by bbaribeau
• Latest reply from bbaribeau
• This topic is resolved

Tags:

No tags yet.

1. 

   bbaribeau
   Member

   I'm trying to create a DSE cluster of 2 search nodes to run the wikipedia sample. Note that everything works fine if I only use one node.

   I can successfully install DSE, OpsCenter, and agents are running on both machines.

   As soon as I run the first script for the wikipedia sample however, my cluster enters a state where only 1 node will ever be up at a time. Restarting one node will cause the other node to crash. They error they produce is:

   ERROR 08:56:19,241 Fatal exception in thread Thread[WRITE-/192.168.62.204,5,main]
   java.lang.StackOverflowError
   at java.net.SocketOutputStream.socketWrite0(Native Method)
   at java.net.SocketOutputStream.socketWrite(Unknown Source)
   at java.net.SocketOutputStream.write(Unknown Source)
   at java.io.BufferedOutputStream.flushBuffer(Unknown Source)
   at java.io.BufferedOutputStream.flush(Unknown Source)
   at java.io.DataOutputStream.flush(Unknown Source)
   at org.apache.cassandra.net.OutboundTcpConnection.writeConnected(OutboundTcpConnection.java:123)
   at org.apache.cassandra.net.OutboundTcpConnection.run(OutboundTcpConnection.java:99)

   The log files for both nodes can be found at: https://docs.google.com/open?id=0B56RRVlR8h2MZENKWjBUN0tsbFE

   Any help as to what's happening would be much appreciated..

   Posted 8 months ago #
2. 

   tjake
   Moderator

   What JVM are you using? There is a issue in the newest Sun 6/7 JVMs requiring more stack. You can set this in conf/cassandra-env.sh

   Most people are 180k now.

   We will update this in the next DSE release.

   Posted 8 months ago #
3. 

   bbaribeau
   Member

   I've followed the steps on: http://www.datastax.com/docs/1.0/install/install_jre#install-jre-rh

   I'm running 1.6 update 34 from Oracle. I've already had to modify cassandra-env.sh to set the minimum stack size to 160k (the default of 128k generates an error when starting dse).

   Also I've installed DSE 2.0 if that makes a difference.

   Posted 8 months ago #
4. 

   sam
   Member

   Going by the system.log files, there seems to be a problem with the schema on one of the nodes. If this is a test of P.O.C. setup, can you nuke the data directory (i.e. /var/lib/cassandra) on both nodes and start over?

   Posted 8 months ago #
5. 

   bbaribeau
   Member

   Finally got around to getting back to this issue.

   Turns out changing the stack size 180k fixed the issue.

   Posted 7 months ago #

RSS feed for this topic

Reply

You must log in to post.