I think your analysis makes sense.
In the else block which is not shown in the snippet, we actually check against null value.
Do you mind filing a Jira ?
On Dec 7, 2011, at 5:30 PM, Shrijeet Paliwal <[EMAIL PROTECTED]> wrote:
> Version in which I observed issue: 0.90.3 + RPC timeout patch
> Version affected : trunk
> Stack trace:
> 2011-11-30 13:10:26,557 [IPC Client (47) connection to
> inw-68.rfiserve.net/172.22.4.68:60020 from an unknown user] WARN
> org.apache.hadoop.ipc.HBaseClient - Unexpected exception receiving call
> I think line number #3 here http://pastie.org/2983633 may return a null
> 'call'. It is so because if you rpc timeout enable, we proactively clean up
> other calls which have expired their lifetime along with the call for which
> timeout exception happend (refer to this method : private void
> cleanupCalls(long rpcTimeout)) .
> Please let me know if what I am thinking makes sense.