Github user vanzin commented on the pull request:
https://github.com/apache/spark/pull/1094#issuecomment-46201382
Hi @rahulsinghaliitd,
Thanks for filing this; this is something I had on my radar but hadn't
gotten down to it yet.
While I also want the feature, I think we need a better approach here. For
one, I'm not a big fan of using the log directory name as the application's URL
in the history server; it's not very discoverable, since it's generated
internally by the event logger. Also, slightly unrelated, but I have upcoming
changes that make those names even more ugly, so they're not a very
user-friendly thing to use as a public-facing URL.
I've actually been playing with this and have some changes that use the
actual application ID, when it's available (such as in the case of Yarn), so
server the application from the history server. I think that's a better
approach. But if some committer wants to check this in while I work on that, it
would cause (many) conflicts with my change.
While we're here, you'd need to make a similar change in the yarn/alpha
project, and also in yarn-client mode (ExecutorLauncher.scala in both projects).
---
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 [email protected] or file a JIRA ticket
with INFRA.
---
