Occasional Infinispan TimeoutException during various OpenShift failover scenarios

There are occasional org.infinispan.util.concurrent.TimeoutException during various failover tests on OpenShift with EAP CD 13 image. The exception sometimes manifests as warning sometimes as error. Sometimes there appears org.infinispan.commons.CacheException: ISPN000482: Cannot create remote transaction GlobalTx.

There are still some HTTP 500 responses from time to time. I've been running two types of tests so far. First is scale down from 3 pods to 2 (with graceful shutdown) and second is docker kill 1 pod of 3 ( followed by Pod restart). Tests are in both variants - using openshift.DNS_PING protocol and/or kubernetes.KUBE_PING protocol. Please see attached logs.