Github user tillrohrmann commented on a diff in the pull request:
https://github.com/apache/flink/pull/3388#discussion_r103195968
--- Diff:
flink-yarn/src/main/java/org/apache/flink/yarn/AbstractYarnClusterDescriptor.java
---
@@ -742,6 +755,24 @@ public FileVisitResult
preVisitDirectory(java.nio.file.Path dir, BasicFileAttrib
}
}
+ for (URI originURI : archives) {
+ Path remoteParent =
Utils.getRemoteResourceRoot(appId.toString(), new Path(originURI.getPath()),
fs.getHomeDirectory());
+ String fragment = originURI.getFragment();
+ Path target = new Path(
+ FileUtils.localizeRemoteFiles(new
org.apache.flink.core.fs.Path(remoteParent.toUri()), originURI).toUri());
+ URI targetURI = target.toUri();
+ if (targetURI.getFragment() == null && fragment !=
null) {
+ targetURI = new URI(target.toUri().toString() +
"#" + fragment);
--- End diff --
Why do we need to append a fragment here? I thought if `remoteParent`
contained a fragment, then `target` would be a file of that fragment and be
named after the fragment.
---
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.
---
