[HDDS-1960] TestMiniChaosOzoneCluster may run until OOME

• Type: Sub-task
• Status: Resolved
• Priority: Major
• Resolution: Duplicate
• Component/s: test
• Labels: MiniOzoneChaosCluster ozone-jira-volunteer ozone-moved-out-critical
• Sprint: HDDS BadLands
• Assignee: Unassigned
• Reporter: Attila Doroszlai
• Created: 13/Aug/19 13:45
• Updated: 09/Oct/23 14:35
• Resolved: 09/Oct/23 14:35

Description

TestMiniChaosOzoneCluster runs load generator on a cluster for supposedly 1 minute, but it may run indefinitely until JVM crashes due to OutOfMemoryError.

In 0.4.1 nightly build it crashed 29/30 times (and no tests were executed in the remaining one run due to some other error).

Latest:
https://github.com/elek/ozone-ci/blob/3f553ed6ad358ba61a302967617de737d7fea01a/byscane/byscane-nightly-wggqd/integration/output.log#L5661-L5662

When it crashes, it leaves GBs of data lying around.

HDDS-1952 disabled this test in CI runs. It can still be run manually (eg. mvn -Phdds -pl :hadoop-ozone-integration-test -Dtest=TestMiniChaosOzoneCluster test). The goal of this task is to investigate the root cause of the runaway nature of this test.

Issue Links

• is a clone of: HDDS-1952 TestMiniChaosOzoneCluster may run until OOME
• is superceded by: HDDS-3131 TestMiniChaosOzoneCluster timeout