[
https://issues.apache.org/jira/browse/HADOOP-17402?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17241893#comment-17241893
]
Steve Loughran commented on HADOOP-17402:
-----------------------------------------
> GCS connector has hadoop.fs.Filesystem service declared like here, is this
> the right way?
yes. Does force a load of the class, which may have side effects or even fail
to load. it can also be very slow. Hence our interest in having some separate
declaration
the S3A error isn't as useful as you think as if the hadoop-aws JAR is present
but the aws-sdk JAR isn't you end up with different linkage problems. If you
look at HADOOP-14132 we've discussed actually including in the service
declaration a list of mandatory dependencies which we could look for and fail
with meaningful messages.
Take a look at my cloudstore JAR to see what I envisage.
https://github.com/steveloughran/cloudstore
> Add GCS FS impl reference to core-default.xml
> ---------------------------------------------
>
> Key: HADOOP-17402
> URL: https://issues.apache.org/jira/browse/HADOOP-17402
> Project: Hadoop Common
> Issue Type: Improvement
> Components: fs
> Reporter: Rafal Wojdyla
> Priority: Major
>
> Akin to current S3 default configuration add GCS configuration, specifically
> to declare the GCS implementation. [GCS
> connector|https://cloud.google.com/dataproc/docs/concepts/connectors/cloud-storage].
> Has this not been done since the GCS connector is not part of the hadoop/ASF
> codebase, or is there any other blocker?
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
