Normen Seemann created CASSANDRA-4596:
-----------------------------------------

             Summary: thrift call to get_paged_slice() hangs if end token is "0"
                 Key: CASSANDRA-4596
                 URL: https://issues.apache.org/jira/browse/CASSANDRA-4596
             Project: Cassandra
          Issue Type: Bug
          Components: API, Core, Hadoop
    Affects Versions: 1.1.3
         Environment: linux
            Reporter: Normen Seemann


I am using get_paged_slice() for range scans driven from within hadoop mappers. 
The mapper that scans the "last segment" with get_paged_slice() where start key 
is set *and* end_token is set to "0" token hangs within the thirft call. Client 
shows following jstack:

 - java.net.SocketInputStream.socketRead0(java.io.FileDescriptor, byte[], int, 
int, int) @bci=0 (Interpreted frame)
 - java.net.SocketInputStream.read(byte[], int, int) @bci=84, line=129 
(Interpreted frame)
 - org.apache.thrift.transport.TIOStreamTransport.read(byte[], int, int) 
@bci=25, line=127 (Interpreted frame)
 - org.apache.thrift.transport.TTransport.readAll(byte[], int, int) @bci=22, 
line=84 (Interpreted frame)
 - org.apache.thrift.transport.TFramedTransport.readFrame() @bci=10, line=129 
(Interpreted frame)
 - org.apache.thrift.transport.TFramedTransport.read(byte[], int, int) @bci=28, 
line=101 (Interpreted frame)
 - org.apache.thrift.transport.TTransport.readAll(byte[], int, int) @bci=22, 
line=84 (Interpreted frame)
 - org.apache.thrift.protocol.TBinaryProtocol.readAll(byte[], int, int) 
@bci=12, line=378 (Interpreted frame)
 - org.apache.thrift.protocol.TBinaryProtocol.readI32() @bci=52, line=297 
(Interpreted frame)
 - org.apache.thrift.protocol.TBinaryProtocol.readMessageBegin() @bci=1, 
line=204 (Interpreted frame)
 - org.apache.thrift.TServiceClient.receiveBase(org.apache.thrift.TBase, 
java.lang.String) @bci=4, line=69 (Interpreted frame)
 - org.apache.cassandra.thrift.Cassandra$Client.recv_get_paged_slice() @bci=12, 
line=727 (Interpreted frame)
 - 
org.apache.cassandra.thrift.Cassandra$Client.get_paged_slice(java.lang.String, 
org.apache.cassandra.thrift.KeyRange, java.nio.ByteBuffer, 
org.apache.cassandra.thrift.ConsistencyLevel) @bci=10, line=711 (Interpreted 
frame)

Changing the end_token from "0" to "2**127-1" fixes the problem, however, I 
would only consider this a workaround. Now, there are actually two issues:

1.) Is the call to get_paged_slice() I described supported at all?
2.) if it is not supported please fix with reasonable error instead of just 
hanging

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Reply via email to