[
https://issues.apache.org/jira/browse/HADOOP-12635?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15092955#comment-15092955
]
Chris Nauroth commented on HADOOP-12635:
----------------------------------------
[~dchickabasapa], I thought of another potential problem with the current patch.
The code I've seen so far is focused on the {{append}} API. I haven't seen any
code changes in the {{create}} code path, so WASB {{create}} calls would not
coordinate on the lease with the {{append}} calls. This means that it would be
possible for process A to open a file for {{create}}, and concurrently process
B opens the same file for {{append}}. Both processes would write data to their
respective streams, and I guess ultimately both could flush and commit blocks
to Azure Storage. This differs from HDFS, which issues a lease on a path for
both {{create}} and {{append}} calls, thus ensuring single-writer semantics
across both {{create}} and {{append}}. In this same example executing against
HDFS, one of the processes would have succeeded, and the other would have
received an error.
If your goal is to match HDFS semantics fully, then the patch would need to do
something about the {{create}} calls too.
> Adding Append API support for WASB
> ----------------------------------
>
> Key: HADOOP-12635
> URL: https://issues.apache.org/jira/browse/HADOOP-12635
> Project: Hadoop Common
> Issue Type: Improvement
> Components: azure
> Affects Versions: 2.7.1
> Reporter: Dushyanth
> Assignee: Dushyanth
> Attachments: Append API.docx, HADOOP-12635.001.patch,
> HADOOP-12635.002.patch
>
>
> Currently the WASB implementation of the HDFS interface does not support
> Append API. This JIRA is added to design and implement the Append API support
> to WASB. The intended support for Append would only support a single writer.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
