Internal error processing get_range_slices

[Patrik Modesto]

Hi,

on our dev cluster of 4 cassandra nodes 0.7.8 I'm suddenly getting:

ERROR 13:40:50,848 Internal error processing get_range_slices
java.lang.OutOfMemoryError: Java heap space
        at java.util.ArrayList.<init>(ArrayList.java:112)
        at 
org.apache.cassandra.service.StorageProxy.getRangeSlice(StorageProxy.java:480)
        at 
org.apache.cassandra.thrift.CassandraServer.get_range_slices(CassandraServer.java:486)
        at 
org.apache.cassandra.thrift.Cassandra$Processor$get_range_slices.process(Cassandra.java:2868)
        at 
org.apache.cassandra.thrift.Cassandra$Processor.process(Cassandra.java:2555)
        at 
org.apache.cassandra.thrift.CustomTThreadPoolServer$WorkerProcess.run(CustomTThreadPoolServer.java:206)
        at 
java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecutor.java:886)
        at 
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:908)
        at java.lang.Thread.run(Thread.java:662)

I run get_range_slices() on all keys with 3 columns named in the Thrift request.

    columnParent.column_family = CATEGORIES_CATEGORY;

    keyRange.start_key         = "";
    keyRange.end_key           = "";
    keyRange.__isset.start_key = true;
    keyRange.__isset.end_key   = true;
    keyRange.count             = std::numeric_limits<int32_t>::max();

    slicePredicate.column_names.push_back(CATEGORIES_CATEGORY_ID);
    slicePredicate.column_names.push_back(CATEGORIES_CATEGORY_NAME);
    slicePredicate.column_names.push_back(CATEGORIES_CATEGORY_PARENT);
    slicePredicate.__isset.column_names = true;

     std::vector<oacassandra::KeySlice>  rangeSlices;
     cassandraWrapper->get_range_slices(rangeSlices, columnParent,
slicePredicate, keyRange, oacassandra::ConsistencyLevel::QUORUM);

There are just 102 rows each with 6 columns. Maximum rowsize is 3 379
391B, mean rowsize is 407 756B. Suddenly Cassandra needs 9GB of
heap-space to fulfill this get_range_slices. There is no cache
enabled.

What could be the problem here?

Regards,
Patrik

PS: while reading the email before I'd send it, I've noticed the
keyRange.count =... is it possible that Cassandra is preallocating
some internal data acording the KeyRange.count parameter?