How to Solve the 'Lock Obtain Timed Out' When Using Solr Plainly

How to solve the 'Lock obtain timed out' when using Solr plainly?

The following modifications solved the issue:

• Applied the changes described at https://stackoverflow.com/a/3035916/236007

• Switched to Oracle Java runtime (it was IBM Java runtime).

• Put the ulimit -v unlimited in /etc/init.d/tomcat7.

• Modified the /usr/share/tomcat7/bin/setenv.sh file as the following (giving it about 4 GB memory):

• 
export JAVA_OPTS="$JAVA_OPTS -Xmx4000m -XX:+HeapDumpOnOutOfMemoryError -XX:HeapDumpPath=/mnt/data/tomcat_dump"


solr multicore post data

Make sure that the core name is in the path:

http://host:8983/solr/CORE/update

When using the post.jar in a multicore setup, you have to override the default url parameter with a url that contains the CORE name:
java -Durl=http://... -jar post.jar

You can type java -jar post.jar -help to get the list of parameters that can be overridden.

SolrException: Internal Server Error

I am very unsure, but in this thread

http://www.mail-archive.com/solr-user@lucene.apache.org/msg08048.html

they recommend to use

<unlockOnStartup>true</unlockOnStartup>

and

<lockType>simple</lockType>

I think this should be safe as long as you access the index through solr or solrj (not though lucene!).

Any other ideas?

---

Related Topics