Steve Loughran commented on HADOOP-14598:

FYI., on a "similar topic but not actually caused by the new work" is the 
discovery that you get an NPE if you try to use an hdfs:// URL For a JAR in the 
classloader: SPARK-21697

That is: you cannot safely use *any* of the Hadoop filesystems as supported 
schemas for classloaders, due to a loop in commons-logging setup & use (at the 
very least, there may be other examples too). Not caused by the recent changes 
to {{FsUrlStreamHandlerFactory}} or the new schemas, just a new codepath that's 
not been looked at before

> Blacklist Http/HttpsFileSystem in FsUrlStreamHandlerFactory
> -----------------------------------------------------------
>                 Key: HADOOP-14598
>                 URL: https://issues.apache.org/jira/browse/HADOOP-14598
>             Project: Hadoop Common
>          Issue Type: Bug
>          Components: fs/azure, test
>    Affects Versions: 2.9.0, 3.0.0-alpha4
>            Reporter: Steve Loughran
>            Assignee: Steve Loughran
>             Fix For: 2.9.0, 3.0.0-beta1
>         Attachments: HADOOP-14598-002.patch, HADOOP-14598-003.patch, 
> HADOOP-14598-004.patch, HADOOP-14598-005.patch
> my downstream-of-spark cloud integration tests (where I haven't been running 
> the azure ones for a while) now have a few of the tests failing
> {code}
>  org.apache.hadoop.fs.azure.AzureException: 
> com.microsoft.azure.storage.StorageException: 
> org.apache.hadoop.fs.FsUrlConnection cannot be cast to 
> java.net.HttpURLConnection
> {code}
> No obvious cause, and it's only apparently happening in some of the 
> (scalatest) tests.

This message was sent by Atlassian JIRA

To unsubscribe, e-mail: common-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: common-issues-h...@hadoop.apache.org

Reply via email to