[
https://issues.apache.org/jira/browse/HADOOP-11550?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Kai Zheng updated HADOOP-11550:
-------------------------------
Description:
In HADOOP-11646, {{ErasureCoder}} API is defined to support encoding/decoding
of a block group considering the most simple but often case that only one
coding step is involved. This is to enhance the API to support complex cases
requiring multiple steps for encoding/decoding of a block group.
Recovering parity blocks will be supported here also, because it involves at
least two steps: step 1. recover data blocks first; step2. call encoder to
compute the erased parity blocks with all the ready data blocks.
was:In HDFS-7662, *ErasureCoder* API is defined to support encoding/decoding
of a block group considering the most simple but often case that only one
coding step is involved. This is to ehance the API to support complex cases
requiring multiple steps for encoding/decoding of a block group.
> Ehance ErasureCoder API to support multiple encoding/decoding steps
> -------------------------------------------------------------------
>
> Key: HADOOP-11550
> URL: https://issues.apache.org/jira/browse/HADOOP-11550
> Project: Hadoop Common
> Issue Type: Improvement
> Reporter: Kai Zheng
> Assignee: Kai Zheng
>
> In HADOOP-11646, {{ErasureCoder}} API is defined to support encoding/decoding
> of a block group considering the most simple but often case that only one
> coding step is involved. This is to enhance the API to support complex cases
> requiring multiple steps for encoding/decoding of a block group.
> Recovering parity blocks will be supported here also, because it involves at
> least two steps: step 1. recover data blocks first; step2. call encoder to
> compute the erased parity blocks with all the ready data blocks.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
