Built the src, and ran the tests using check-all.sh. Most of the tests ran
fine. There was an transient test failure (
https://issues.apache.org/jira/browse/SAMZA-1405), which seems to be caused
by the testing env (further investigation needed). I reran the tests again
and it passed. Since this tes
GitHub user bharathkk opened a pull request:
https://github.com/apache/samza/pull/281
SAMZA-1404: Add warning in case of potential process starvation due to
longer window method
Currently, we use the average windowNs as the lower bound for window
trigger time to determine if user n
Downloaded the source, compiled and ran the integration tests. All passed.
+1 (binding) w/ the following minor comments:
# Please make a note in the release note that this version requires JDK
1.8.0.112+ (I have test w/ JDK 1.8.0.121)
# Please make sure that we publish artifacts compiled w/ Scala
Hi Jeremiah,
Can you check if your classpath has multiple SLF4J bindings and if it's
picking something other than log4j (e.g. logback) instead? If that's the
case, the other implementation might not have an explicit configuration and
default to logging everything to stdout.
You can verify by chec
I am seeing tons of noise in container logs after upgrading to samza 0.13.
Changes to my job's log4j properties aren't affecting this output. Also, logs
are now pointing output to stdout instead of the samza-cotnainer-n.log file.
18:42:19.761 [SAMZA-JVM-METRICS] DEBUG org.apache.samza.metrics.J
