zentol commented on a change in pull request #11372: [FLINK-16480] Improve
error reporting on AZP
URL: https://github.com/apache/flink/pull/11372#discussion_r393598579
##########
File path: tools/travis_watchdog.sh
##########
@@ -177,6 +177,14 @@ print_stacktraces () {
done
}
+collect_coredumps() {
+ echo "Searching for .dump, .dumpstream and related files in
$($HERE/../)"
+ for file in `find . -type f -regextype posix-extended -iregex
'.*\.dump|.*\.dumpstream|.*hs.*\.log|.*/core(.[0-9]+)?$'`; do
Review comment:
> But who guarantees that Kafka, ElasticSearch, Zookeeper all properly
respect the environment variables.
There's no guarantee that these are run under FLINK_ROOT though; the java
e2e tests use a junit temporary folder to store and run these services, as well
as the flink distribution.
At least you'd also have to search the entire tmp directory of the system,
or modify `java.io.tmpdir` to point somewhere under FLINK_ROOT.
Have you tried the `JAVA_TOOL_OPTIONS` environment variable that the JVM
itself evaluates?
https://docs.oracle.com/javase/8/docs/technotes/guides/troubleshoot/envvars002.html
> I don't know if I would spend my time well for the project if I would
refactor something that works.
This can be read in rather frightening ways: "Lol let's use the first
version that works _now_."
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
