Stopping and starting all LPARS deployed on power servers for SAP App and DB

From Notes_Wiki

Home > Suse > SAP setup and maintenance > Stopping and starting all LPARS deployed on power servers for SAP App and DB

It is possible to required shutting down all LPARs, followed by power server for some UPS / Cooling / Network maintenance activity at DC. In such cases to shutdown and start entire SAP related infrastructure refer below steps:


pacemaker should not auto-start on node after reboot

The below article assumed paceamaker service does not automatically start on App and DB servers on OS boot. If the App or DB has pacemaker service enabled to automatically start on boot then steps might differ. Especially in that case the primary App (ACS) might start on secondary node and ERS might start on primary node. Hence best option is to have auto-start of pacemaker disabled on machie boot.

If that is not possible, we should consider putting secondary node in maintenance mode before removing cluster from maintenance mode using:

crm node maintenance <NODENAME>

Once cluster services are up and running on primary properly. Then for secondary node use:

crm node ready NODENAME


SAP Infra shutdown process

  1. Put both application and DB clusters in maintenance mode using:
    crm configure property maintenance-mode=true
    See Pacemaker start and stop for Hana DB and application servers
    No point in stopping pacemaker service as anyway we are going to shutdown the LPARs
  2. Stop SAP Application and then SAP DB
  3. We can use vMHC to shutdown all LPARS using graphical shutdown option. After selecting shutdown choose "Operating System" option. There are also option such as "Delayed", "Immediate" or "Operating System Immediate".
    Refer: https://www.redbooks.ibm.com/redbooks/pdfs/sg247491.pdf
  4. Or if vHMC based shutdown is not preferred, shutdown all the LPAR's using:
    shutdown -h now
  5. Shutdown second VIOS server
    In vHMC after clicking on a power server we get options such as LPARS and "VIOS Servers". LPARS are visible under LPARS and "VIOS servers" are visible under VIOS server. Same as for LPAR choose action "Shutdown" with option "Operating System"
  6. Shutdown first VIOS server
  7. Shutdown second physical host
    This can be done from vHMC. There is option to poweroff the servers.
  8. Shutdown first physical host
  9. Shutdown HMC VM
    You might have to poweroff (unclean shutdown) of VM from ESXi. Still need to figure out how to shutdown vHMC VM properly.
  10. Shutdown Physical server hosting HMC VM
  11. Shutdown storage
    In a few storages option to shutdown can be part of Monitoring tab.
  12. Shutdown Tape Library
    There might be option to change "Library State" to Offline, if there is no shutdown option.


SAP Infra start up procedure

  1. Start Tape Library
    Might have to change Library state to Online
  2. Start storage
    In a few storages there is only option to shutdown. There is no power on button. As soon as A/C power is restored the storage might get powered on automatically.
  3. Start host where vHMC VM is deployed. This can be done using OOBE such iDRAC, IMM or iLO, etc.
  4. Start HMC VM after ESXi has booted properly.
    Note: After storage, wait for 15-20mins for start physical hosts
  5. Start primary physical host
  6. Start seconday physical host
    After starting both the status of one host might be "Operating" and other might be "Standby". That is fine. After this we can proceed with starting the VIOS.
  7. Start primary VIOS server (Activate -> Normal)
  8. Start secondary VIOS server (Activate -> Normal)
  9. Start all LPARs on primary host (Activate -> Normal)
  10. Start all LPARs on secondary host (Activate -> Normal)
  11. SAP team should start SAP DB on primary server.
    See Pacemaker start and stop for Hana DB and application servers
  12. Also SAP application team should confirm that primary to secondary DB replication is in proper state
  13. Start pacemaker services on primary DB host and remove cluster from maintenance mode.
    systemctl start pacemaker.service
    crm configure property maintenance-mode=false
    crm status
  14. After 5 minutes start pacemaker services on secondary DB host using:
    systemctl start pacemaker.service
    crm status
  15. After DB is successfully up along with primary to secondary replication, Start App cluster on primary host:
    systemctl start pacemaker.service
    crm configure property maintenance-mode=false
    crm status
  16. After this start pacemaker cluster on secondary App host after waiting for 5 min
    systemctl start pacemaker.service
    crm status
    Secondary App host might reboots automatically before becoming part of cluster.
  17. After this SAP application team can start other APP services on App nodes

Refer:


Home > Suse > SAP setup and maintenance > Stopping and starting all LPARS deployed on power servers for SAP App and DB