[HADOOP-4802] RPC Server send buffer retains size of largest response ever sent

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Duplicate
• Affects Version/s: 0.18.2, 0.19.0
• Fix Version/s: 0.20.2
• Component/s: ipc
• Assignee: Unassigned
• Reporter: Michael Stack
• Created: 07/Dec/08 19:45
• Updated: 02/Jul/10 04:46
• Resolved: 04/Jan/10 07:37

Description

The stack-based ByteArrayOutputStream in Server.Hander is reset each time through the run loop. This will set the BAOS 'size' back to zero but the allocated backing buffer is unaltered. If during an Handlers' lifecycle, any particular RPC response was fat – Megabytes, even – the buffer expands during the write to accommodate the particular response but then never shrinks subsequently. If a hosting Server has had more than one 'fat payload' occurrence, the resultant occupied heap can provoke memory woes (See https://issues.apache.org/jira/browse/HBASE-900?focusedCommentId=12654009#action_12654009 for an extreme example; occasional payloads of 20-50MB with 30 handlers robbed the heap of 700MB).

Issue Links

• duplicates: HADOOP-6460 Namenode runs of out of memory due to memory leak in ipc Server
• is related to: HADOOP-4813 Avoid a buffer copy while replying to RPC requests.