Can't understand the problem from the log

I'm trying to run

$ esrally --track-path=track.json --preserve-install true --car="4gheap" --client-options="timeout:6000" --distribution-version=7.0.0-rc2

Here's what I have in log https://paste.apache.org/0wRo
Something happen, but I can't grasp what.
Elastic log is absent. Then when I'me trying to launch elastic I've got a culprit.

:~/.rally/benchmarks/races/2019-04-09-12-22-24/rally-node-0/install/elasticsearch-7.0.0-rc2/bin$ ./elasticsearch --verbose
OpenJDK 64-Bit Server VM warning: Option UseConcMarkSweepGC was deprecated in version 9.0 and will likely be removed in a future release.
ERROR: Temporary file directory [${ES_TMPDIR}] does not exist or is not accessible

Here's my java

:~/.rally/benchmarks/races/2019-04-09-12-22-24/rally-node-0/install/elasticsearch-7.0.0-rc2/bin$ $JAVA_HOME/bin/java --version
openjdk 12 2019-03-19
OpenJDK Runtime Environment AdoptOpenJDK (build 12+33)
OpenJDK 64-Bit Server VM AdoptOpenJDK (build 12+33, mixed mode, sharing)

Checking tmp dir write access.

:~/.rally/benchmarks/races/2019-04-09-12-22-24/rally-node-0/install/elasticsearch-7.0.0-rc2/bin$ cat > ${ES_TMPDIR}/o
sss

:~/.rally/benchmarks/races/2019-04-09-12-22-24/rally-node-0/install/elasticsearch-7.0.0-rc2/bin$ file ${ES_TMPDIR}/o
/tmp/o: ASCII text, with no line terminators

How I can understand what's going on from rally logs? Do you know how to deal with that ES_TMPDIR failure?

Hi Mikhail,

Unfortunately you are hitting: https://github.com/elastic/elasticsearch/issues/40812

Rally picks up the configuration option from the rally-teams repo (7 branch) which contains the right content for -Djava.io.tmpdir (without double quotes) for the latest 7.0.0. The 7.0.0-rc2 artifact, that Rally downloads, requires double quotes for this setting.

Since you are having a local setup of Rally and Elasticsearch you could also switch to the from-sources-complete pipeline and pick one of the commits of the Elasticsearch 7.0 branch using --revision= until 7.0.0 is GA. This way Rally will build Elasticsearch locally (just make sure you have the right JAVAx_HOME env vars set see also here -- for Elasticsearch 7 you'll need JAVA12_HOME to build and if you want to use a different jvm to run Elasticsearch with, set JAVAx_HOME accordingly and specify the major version using --runtime-jdk).

Regards,
Dimitris

Ok. Thanks, @dliappis. It clarifies the issue. But how we can make it easier understandable from rally.log? Can we make it log stderr from elastic?

This should be present in rally.log. We reproduced this using:

esrally --distribution-version=7.0.0-rc2 --test-mode --challenge=append-no-conflicts-index-only

and could see the following in rally.log

2019-04-10 10:03:51,772 ActorAddr-(T|:52550)/PID:41926 esrally.mechanic.launcher INFO Allowed JDK versions are [12, 11, 8].
2019-04-10 10:03:51,949 ActorAddr-(T|:52550)/PID:41926 esrally.mechanic.launcher INFO Detected JDK with major version [12] in [/Library/Java/JavaVirtualMachines/jdk-12.jdk/Contents/Home].
2019-04-10 10:03:51,949 ActorAddr-(T|:52550)/PID:41926 esrally.mechanic.launcher INFO Starting node [rally-node-0] based on car [defaults].
2019-04-10 10:03:52,3 ActorAddr-(T|:52550)/PID:41926 esrally.utils.process INFO b'openjdk version "12-ea" 2019-03-19\n'
2019-04-10 10:03:52,4 ActorAddr-(T|:52550)/PID:41926 esrally.utils.process INFO b'OpenJDK Runtime Environment (build 12-ea+24)\n'
2019-04-10 10:03:52,4 ActorAddr-(T|:52550)/PID:41926 esrally.utils.process INFO b'OpenJDK 64-Bit Server VM (build 12-ea+24, mixed mode, sharing)\n'
2019-04-10 10:03:52,8 ActorAddr-(T|:52550)/PID:41926 esrally.mechanic.launcher INFO Setting [-XX:+ExitOnOutOfMemoryError] to detect out of memory errors during the benchmark.
2019-04-10 10:03:52,409 ActorAddr-(T|:52550)/PID:41926 esrally.mechanic.launcher INFO rally-node-0 (stdout): OpenJDK 64-Bit Server VM warning: Option UseConcMarkSweepGC was deprecated in version 9.0 and will likely be removed in a future release.
2019-04-10 10:03:53,488 ActorAddr-(T|:52550)/PID:41926 esrally.mechanic.launcher INFO rally-node-0 (stdout): ERROR: Temporary file directory [${ES_TMPDIR}] does not exist or is not accessible
2019-04-10 10:03:53,570 ActorAddr-(T|:52550)/PID:41926 esrally.mechanic.launcher INFO rally-node-0 (stdout): No more output. Process has likely terminated.
2019-04-10 10:03:53,675 ActorAddr-(T|:52550)/PID:41926 esrally.mechanic.launcher ERROR Node [rally-node-0] has terminated with exit code [78].
2019-04-10 10:03:53,697 ActorAddr-(T|:52524)/PID:41787 esrally.actor INFO Received a benchmark failure from [ActorAddr-(T|:52537)] and will forward it now.

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.