Using Hbase backend, got org.apache.hadoop.hbase.client.ScannerTimeoutException

[Christine Li]

My graph has 295 vertices and 1000 edges. A long query works fine in gremlin console. Rewrite it in Java and got the following error:
java.lang.RuntimeException: org.apache.hadoop.hbase.client.ScannerTimeoutException: 60780ms passed since the last invocation, timeout is currently set to 60000
    at org.apache.hadoop.hbase.client.AbstractClientScanner$1.hasNext(AbstractClientScanner.java:94)
    at com.google.common.collect.Iterators$7.computeNext(Iterators.java:650)
    at com.google.common.collect.AbstractIterator.tryToComputeNext(AbstractIterator.java:143)
    at com.google.common.collect.AbstractIterator.hasNext(AbstractIterator.java:138)
    at com.thinkaurelius.titan.diskstorage.hbase.HBaseKeyColumnValueStore$RowIterator.hasNext(HBaseKeyColumnValueStore.java:277)
    at com.thinkaurelius.titan.graphdb.database.StandardTitanGraph$2.hasNext(StandardTitanGraph.java:391)
    at com.thinkaurelius.titan.graphdb.transaction.VertexIterable$1.nextVertex(VertexIterable.java:34)
    at com.thinkaurelius.titan.graphdb.transaction.VertexIterable$1.next(VertexIterable.java:55)
    at com.thinkaurelius.titan.graphdb.transaction.VertexIterable$1.next(VertexIterable.java:27)

I I am using the same properties files for both cases:
cache.db-cache = true
cache.db-cache-clean-wait = 20
cache.db-cache-time = 180000
cache.db-cache-size = 0.5

And hbase-site.xml has:

    <property>
      <name>hbase.client.scanner.caching</name>
      <value>100</value>
    </property>

I don't quite understand why it works ok in groovy, but doesn't work in Java. And Should I reduce the scanner caching to resolve the problem?

Thanks,

[anton...@gmail.com]

The exception is saying that you have exceeded the timeout set namely 60780ms while the configured limit is 60000ms

can you check if these settings discussed here would help

http://stackoverflow.com/questions/11463458/hbase-mapreduce-error

[Christine Li]

I've got the problem resolved. Apparently, I need to modify the hbase setting from hbase library on my local machine. In my case, HBase is on a remote machine and used as the storage backend for Titan graph database.

I found that hbase-default.xml was built inside the hbase-common jar. When a query is executed from my application, the settings from the hbase-default.xml are used. Not what I thought that the settings are retrieved from the hbase server.

I suppose the setting on hbase server will work for queries executed from hbase shell.