[ https://issues.apache.org/jira/browse/GEODE-8503?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
ASF GitHub Bot updated GEODE-8503: ---------------------------------- Labels: pull-request-available (was: ) > Limit netty threads used in PubSubDUnitTest > ------------------------------------------- > > Key: GEODE-8503 > URL: https://issues.apache.org/jira/browse/GEODE-8503 > Project: Geode > Issue Type: Test > Components: redis > Reporter: Jens Deppe > Assignee: Jens Deppe > Priority: Major > Labels: pull-request-available > > The {{PubSubDUnitTest}} currently fails when run under stress > ({{StressNewTest}} job in CI). There are various factors at play here: > - Test hosts have 96 vCPUs and netty will use 2x this number for each > EventLoopGroup (of which we have 2). For some tests we end up needing almost > 1600 threads which seems excessive. > - VMs are not recycled between every test run but thread resources are also > not cleaned up synchronously between runs; this seems to be the biggest > problem. -- This message was sent by Atlassian Jira (v8.3.4#803005)