[
https://issues.apache.org/jira/browse/HDFS-8481?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Zhe Zhang updated HDFS-8481:
----------------------------
Attachment: HDFS-8481-HDFS-7285.04.patch
Thanks Kai for verifying this. I'm attaching 04 patch to address minor issues
above.
To address the GC issue we should also avoid filling 0 bytes. Maybe the codec
can support a special flag to mark an input slot as all-zero?
I'm currently working on reusing the input/output buffers. It turns out tricky
because 1) we need to change all byte arrays to {{ByteBuffer} and 2) we need a
better abstraction to divide the rounds of {{decode()}} aligned at cell
boundaries. Perhaps something like a {{StripedDecoder}}.
If the 04 patch looks OK for removing decoding workaround, how about we commit
it first while we work on the various tasks discussed above to reuse all input
and output buffers?
> Erasure coding: remove workarounds in client side stripped blocks recovering
> ----------------------------------------------------------------------------
>
> Key: HDFS-8481
> URL: https://issues.apache.org/jira/browse/HDFS-8481
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Reporter: Zhe Zhang
> Assignee: Zhe Zhang
> Attachments: HDFS-8481-HDFS-7285.00.patch,
> HDFS-8481-HDFS-7285.01.patch, HDFS-8481-HDFS-7285.02.patch,
> HDFS-8481-HDFS-7285.03.patch, HDFS-8481-HDFS-7285.04.patch
>
>
> After HADOOP-11847 and related fixes, we should be able to properly calculate
> decoded contents.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
