[
https://issues.apache.org/jira/browse/HDFS-9643?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15107693#comment-15107693
]
Bob Hansen commented on HDFS-9643:
----------------------------------
That is looking very good.
When running on a TCP connection (rather than in unit tests), when a request is
hanging because the server has gone catatonic (say... thrashing in GC), how
does the current patch tickle the pipeline to check the cancel flag and bail
out? I assumed that the only way to do that reliably was to close the tcp
connection when the handle is cancelled (see
http://www.boost.org/doc/libs/1_60_0/doc/html/boost_asio/reference/basic_stream_socket/cancel/overload2.html),
and I don't see that in this patch. Is there something else going on that I
have missed?
> libhdfs++: Support async cancellation of read operations
> --------------------------------------------------------
>
> Key: HDFS-9643
> URL: https://issues.apache.org/jira/browse/HDFS-9643
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: hdfs-client
> Reporter: James Clampffer
> Assignee: James Clampffer
> Attachments: HDFS-9643.HDFS-8707.000.patch,
> HDFS-9643.HDFS-8707.001.patch
>
>
> It should be possible for any thread to cancel operations in progress on a
> FileHandle. Any ephemeral objects created by the FileHandle should free
> resources as quickly as possible.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
