ben-roling commented on a change in pull request #666: HADOOP-16221 add option
to fail operation on metadata write failure
URL: https://github.com/apache/hadoop/pull/666#discussion_r276058353
##########
File path:
hadoop-common-project/hadoop-common/src/main/resources/core-default.xml
##########
@@ -1522,6 +1522,19 @@
</description>
</property>
+<property>
+ <name>fs.s3a.metadatastore.fail.on.write.error</name>
Review comment:
I've been looking at it a little differently. Primarily, I've been
struggling with whether anyone would really _ever_ want this setting off
(authoritative mode or not). Why bother using S3Guard if you're ok with the
idea that it won't even necessarily prevent seeing list-after-create
inconsistencies? You're effectively getting the same behavior as if you didn't
use S3Guard - most of the time you see all the files and sometimes you don't.
I considered crafting this patch without there even being a configuration
for this or with the default to true, but figured current behavior must have
been intentional and there is an argument I'm not seeing. Can you guys help me
see the side of this where someone would want this false?
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
