[
https://issues.apache.org/jira/browse/HADOOP-1470?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12508917
]
Raghu Angadi commented on HADOOP-1470:
--------------------------------------
Before the above proposal, 'large buffer size' was passed on to Implementation.
Now, looks like implementation does not (or should not) buffer the data it
reads (usually using BufferedInputStream) and should try to read multiple
chunks (without extra copy).
DFS can avoid only one of the two : one socket.read() per chunk or extra copy.
It can not avoid both. If InputChecker does not manage the larger buffer also,
rather just passes it on to implementation, then it can avoid both. But cost of
socket.read() per chunk is probably negligible.
> Rework FSInputChecker and FSOutputSummer to support checksum code sharing
> between ChecksumFileSystem and block level crc dfs
> ----------------------------------------------------------------------------------------------------------------------------
>
> Key: HADOOP-1470
> URL: https://issues.apache.org/jira/browse/HADOOP-1470
> Project: Hadoop
> Issue Type: Improvement
> Components: fs
> Affects Versions: 0.12.3
> Reporter: Hairong Kuang
> Assignee: Hairong Kuang
> Fix For: 0.14.0
>
> Attachments: GenericChecksum.patch, genericChecksum.patch,
> GenericChecksum1.patch, InputChecker-01.java
>
>
> Comment from Doug in HADOOP-1134:
> I'd prefer it if the CRC code could be shared with CheckSumFileSystem. In
> particular, it seems to me that FSInputChecker and FSOutputSummer could be
> extended to support pluggable sources and sinks for checksums, respectively,
> and DFSDataInputStream and DFSDataOutputStream could use these. Advantages of
> this are: (a) single implementation of checksum logic to debug and maintain;
> (b) keeps checksumming as close to possible to data generation and use. This
> patch computes checksums after data has been buffered, and validates them
> before it is buffered. We sometimes use large buffers and would like to guard
> against in-memory errors. The current checksum code catches a lot of such
> errors. So we should compute checksums after minimal buffering (just
> bytesPerChecksum, ideally) and validate them at the last possible moment
> (e.g., through the use of a small final buffer with a larger buffer behind
> it). I do not think this will significantly affect performance, and data
> integrity is a high priority.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
