[
https://issues.apache.org/jira/browse/HADOOP-4379?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12710546#action_12710546
]
stack commented on HADOOP-4379:
-------------------------------
v9 works well in basic testing. If I crash the writer application, a
subsequent reader can pick up edits from the crashed writer if the reader
process does append, close and then reopen.
I've been trying to break the patch but it seems robust (v7 and v8 had issues
that Dhruba fixed).
The append/close process can take from 10-60 seconds in my basic testing.
Usually its about 20 seconds. It would be better if there were no pause. Can
this be addressed?
> In HDFS, sync() not yet guarantees data available to the new readers
> --------------------------------------------------------------------
>
> Key: HADOOP-4379
> URL: https://issues.apache.org/jira/browse/HADOOP-4379
> Project: Hadoop Core
> Issue Type: New Feature
> Components: dfs
> Reporter: Tsz Wo (Nicholas), SZE
> Assignee: dhruba borthakur
> Priority: Blocker
> Fix For: 0.19.2
>
> Attachments: 4379_20081010TC3.java, fsyncConcurrentReaders.txt,
> fsyncConcurrentReaders3.patch, fsyncConcurrentReaders4.patch,
> fsyncConcurrentReaders5.txt, fsyncConcurrentReaders6.patch,
> fsyncConcurrentReaders9.patch, hypertable-namenode.log.gz, namenode.log,
> namenode.log, Reader.java, Reader.java, reopen_test.sh, ReopenProblem.java,
> Writer.java, Writer.java
>
>
> In the append design doc
> (https://issues.apache.org/jira/secure/attachment/12370562/Appends.doc), it
> says
> * A reader is guaranteed to be able to read data that was 'flushed' before
> the reader opened the file
> However, this feature is not yet implemented. Note that the operation
> 'flushed' is now called "sync".
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
