[
https://issues.apache.org/jira/browse/HBASE-8498?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13652717#comment-13652717
]
Anoop Sam John commented on HBASE-8498:
---------------------------------------
>But why did using read() fail when we are using Protobuf WAL and not during
>the Sequential WAL
In case of SequenceFileLogReader we make use of DataInputStream.readInt() which
internally calls is.read 4 times (byte by byte) So it will guarentee to read 4
bytes (If not real EOF)
In case of FSInputStream.read(byte[]) it may not fill the byte[] not only in
real EOF.(This is what we have seen) Do there some clear documentation from
HDFS on when all they will not read and fill the byte[]?
+1
> PB WAL reading is broken due to some partial reads
> --------------------------------------------------
>
> Key: HBASE-8498
> URL: https://issues.apache.org/jira/browse/HBASE-8498
> Project: HBase
> Issue Type: Bug
> Reporter: Sergey Shelukhin
> Assignee: Sergey Shelukhin
> Priority: Critical
> Attachments: HBASE-8498-v0.patch, HBASE-8498-v0.patch,
> HBASE-8498-v1.patch, HBASE-8498-v2.patch, HBASE-8498-v4.patch
>
>
> From dev list, there's IOException during log splitting on hadoop-2.0.3. We
> need to understand whether this is caused by EOF, and ignore it if it is.
--
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
