[prev in list] [next in list] [prev in thread] [next in thread] 

List:       hadoop-dev
Subject:    [jira] [Created] (HADOOP-17498) Add a MkdirOperation for chained S3 operations during mkdir
From:       "Steve Loughran (Jira)" <jira () apache ! org>
Date:       2021-01-26 16:19:01
Message-ID: JIRA.13354713.1611677898000.516278.1611677941396 () Atlassian ! JIRA
[Download RAW message or body]

Steve Loughran created HADOOP-17498:
---------------------------------------

             Summary: Add a MkdirOperation for chained S3 operations during mkdir
                 Key: HADOOP-17498
                 URL: https://issues.apache.org/jira/browse/HADOOP-17498
             Project: Hadoop Common
          Issue Type: Sub-task
          Components: fs/s3
    Affects Versions: 3.3.1
            Reporter: Steve Loughran



on S3A mkdirs is implemented as a walk up the tree looking for any parent dir, \
exiting fast when found; failing if a file is found.

Proposed: pull this out into its own MkDir operation with the goals of efficiency, \
track duration and isolate/test

* specific callbacks to getFileStatus and  createFakeDirectory(key);
* Move from simple getFileStatus call to probes assuming dir exists before looking \
                for a file: HEAD + path /, LIST path,
* the probe for a file  HEAD path will only succeed in a failure mode where higher \
cost is a detail

I think I'd also like to start looking at what we can do with some context \
tracking/telemetry where something provides an operation ID and ability to attach \
this to the requests



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

---------------------------------------------------------------------
To unsubscribe, e-mail: common-dev-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-dev-help@hadoop.apache.org


[prev in list] [next in list] [prev in thread] [next in thread]