Github user jerryshao commented on the issue:
https://github.com/apache/spark/pull/16432
Thanks @tgravescs for your comments. Yes actually it should be Hadoop
compatible file systems, or in other words should be Hadoop `FileSystem` API
supported FS. I tested with kerberized HDFS and wasb.
In HDI, it ships with kerberized wasb or ADSL, they will also have
delegation token but with different token kind. In the current code we assume
that token kind should be `HDFS_DELEGATION_KIND`, this is not true if the
default FS is not HDFS or there's no HDFS in the cluster. In such case
currently we simply ignore the tokens from these FS, which will be an issue for
long running Spark application. This also applies to webHDFS (it also has
different token kind).
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
