[
https://issues.apache.org/jira/browse/HADOOP-14144?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Sean Mackrory updated HADOOP-14144:
-----------------------------------
Attachment: HADOOP-14144-HADOOP-13345.001.patch
Attaching a patch that eliminates (and tests for) duplicates in the output,
previously caused by printDiff calls being scattered somewhat inconsistently. I
also did a general / audit clean up of the logic as there were a few things
that seemed off and implemented the "TODO" on comparing the internal contents
of the files. I ignore mod_time on directories since S3 just reports the
current time for that, and I ignore block_size because I would imagine that
difference would be exceptionally rare and displaying that would just clutter
things (just realized I never added modification time to the output - any
opinions on doing / not doing that? Would need to add something to the UI to
make it's meaning clearer).
> s3guard: CLI diff non-empty after import on new table
> -----------------------------------------------------
>
> Key: HADOOP-14144
> URL: https://issues.apache.org/jira/browse/HADOOP-14144
> Project: Hadoop Common
> Issue Type: Sub-task
> Components: fs/s3
> Reporter: Aaron Fabbri
> Assignee: Sean Mackrory
> Priority: Minor
> Attachments: HADOOP-14144-HADOOP-13345.001.patch
>
>
> I expected the following steps to yield zero diff from `hadoop s3guard diff`
> command.
> (1) hadoop s3guard init ... (create fresh table)
> (2) hadoop s3guard import (fresh table, existing bucket with data in it)
> (3) hadoop s3guard diff ..
> Instead I still get a non-zero diff on step #3. I also noticed some entries
> are printed twice.
> {noformat}
> dude@computer:~/Code/hadoop$ hadoop s3guard diff -meta dynamodb://dude-dev
> -region us-west-2 s3a://dude-dev
> S3 D s3a://dude-dev/user/fabbri/test/parentdirdest
> S3 D s3a://dude-dev/user/fabbri/test/parentdirdest
> {noformat}
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
