Fixed handling of force removal of stuck podman containers in stopping state

Author: tanmaygarg-oracle

OracleIdentityGovernance/samples/scripts/agentManagement.sh

@@ -718,6 +718,27 @@ else
 fi
 }
 
+forceRmPodman()
+{
+# Get the main process for the container.
+CONTAINER_ID=$(podman ps -a | grep "$AI" | awk '{print $1}')
+if [ -n "${CONTAINER_ID}" ]; then
+   echo Container ID : "$CONTAINER_ID"
+   CONTAINER_PROCESS_ID=$(ps -ef | grep -v grep | grep "$CONTAINER_ID" | awk '{print $2}')
+   echo Container Process ID: ${CONTAINER_PROCESS_ID}
+
+   # Kill any processes containing the process ID.
+   # This kills the child processes too.
+   kill -9 `ps -ef | grep -v grep | grep ${CONTAINER_PROCESS_ID} | awk '{print $2}'`
+
+   # Stop the container, as Podman doesn't notice the processes are dead until you interact with the container.
+   echo "Removing container. Ignore errors."
+   podman rm -f "$AI"
+else
+   echo "Container Already Removed"
+fi
+}
+
 # shellcheck source=/dev/null
 stop()
 {
@@ -779,6 +800,7 @@ kill()
         podman exec "$AI" /bin/bash -c 'agent --config /app/data/conf/config.json ido lcm -i status_check; while [[ "$?" != "2" && "$?" != "255" ]]; do sleep 5s;agent --config /app/data/conf/config.json ido lcm -i status_check; done' >/dev/null
       fi
       podman rm -f "$AI"
+      forceRmPodman
   fi
 }