+++ This bug was initially created as a clone of Bug #2155453 +++ Description of problem: The current version of the fence agent fence_ibm_powervs uses --method=onoff for the action reboot. The current version leads to a twice as long failover time compared to the new proposal. Version-Release number of selected component (if applicable): File: https://github.com/ClusterLabs/fence-agents/blob/main/agents/ibm_powervs/fence_ibm_powervs.py Change Date: Oct 25 2022 Tested Version: https://github.com/ClusterLabs/fence-agents/blob/3373431dc49d6e429bbf613765385cb33a56e917/agents/ibm_powervs/fence_ibm_powervs.py How reproducible: always Steps to Reproduce: 1. Deploy two PowerVS LPARs for SAP HANA using RHEL8.4 image see https://cloud.ibm.com/docs/power-iaas?topic=power-iaas-creating-power-virtual-server and https://cloud.ibm.com/docs/sap?topic=sap-hana-iaas-offerings-profiles-power-vs 2. Install SAP HANA 2.0 on both nodes see https://help.sap.com/docs/SAP_HANA_PLATFORM/2c1988d620e04368aa4103bf26f17727/7eb0167eb35e4e2885415205b8383584.html?locale=en-US 3. Setup SAP HANA System Replication and its pacemaker HSR cluster policy see https://access.redhat.com/articles/3004101 4. Create PowerVS fence agent with command: pcs stonith create fence_device fence_ibm_powervs token=${APIKEY} crn=${IBMCLOUD_CRN} instance=${GUID} region=${CLOUD_REGION} api-type=private proxy=http://${PROXY_IP}:3128 pcmk_host_map="${NODE1}:${POWERVSI_01};${NODE2}:${POWERVSI_02}" pcmk_reboot_timeout=600 pcmk_monitor_timeout=600 5. Delay reboot with GRUB_TIMEOUT=3600 in /etc/default/grub and “grub2-mkconfig -o /boot/grub2/grub.cfg” (set pw for console user before reboot) 6. reboot primary HSR node using “sync; echo b > /proc/sysrq-trigger” Actual results: Messages like the following will be shown: Node List: * Node sap-ha-s1-1: online: * Resources: * fence_device (stonith:fence_ibm_powervs): Started (Monitoring) * SAPHanaTopology_ASD_00 (ocf::heartbeat:SAPHanaTopology): Started * SAPHana_ASD_00 (ocf::heartbeat:SAPHana): Slave * Node sap-ha-s1-2: UNCLEAN (offline): * Resources: * fence_device (stonith:fence_ibm_powervs): Started (Monitoring) * SAPHanaTopology_ASD_00 (ocf::heartbeat:SAPHanaTopology): Started * vip_ASD_00 (ocf::heartbeat:IPaddr2): Started * SAPHana_ASD_00 (ocf::heartbeat:SAPHana): Master -> Node sap-ha-s1-2 is flagged UNCLEAN and Node sap-ha-s1-1 does takeover after min. 2 minutes. Expected results: -> Node sap-ha-s1-2 should be flagged (offline) and Node sap-ha-s1-1 should takeover VIP and HANA Primary in less than 2 minutes (Most cases 1 minute). Additional info: https://github.com/ClusterLabs/fence-agents/pull/542