Describe désert/SAP/HANA-Maintenance here.

Migrating and SAP HANA Primary

HDB资源之主备切换

Pre move

• Set the rsc_SAPHana_SLE_HDB00 to be maintenance.

~# crm resouce maintenance rsc_SAPHana_SLE_HDB00

saprep01:~ # crm status detail
Stack: corosync
Current DC: saprep01 (1) (version 1.1.15-19.15-e174ec8) - partition with quorum
Last updated: Tue Aug  8 16:02:14 2023
Last change: Tue Aug  8 15:52:13 2023 by root via crm_resource on saprep01

2 nodes configured
6 resources configured

Online: [ saprep01 (1) saprep02 (2) ]

 stonith-sbd    (stonith:external/sbd): Started saprep01
 rsc_ip_SLE_HDB00       (ocf::heartbeat:IPaddr2):       Started saprep02
 Master/Slave Set: msl_SAPHana_SLE_HDB00 [rsc_SAPHana_SLE_HDB00] (unmanaged)
     rsc_SAPHana_SLE_HDB00      (ocf::suse:SAPHana):    Slave saprep01 (unmanaged)
     rsc_SAPHana_SLE_HDB00      (ocf::suse:SAPHana):    Master saprep02 (unmanaged)
 Clone Set: cln_SAPHanaTopology_SLE_HDB00 [rsc_SAPHanaTopology_SLE_HDB00]
     rsc_SAPHanaTopology_SLE_HDB00      (ocf::suse:SAPHanaTopology):    Started saprep01
     rsc_SAPHanaTopology_SLE_HDB00      (ocf::suse:SAPHanaTopology):    Started saprep02
     Started: [ saprep01 saprep02 ]

Manual Takeove Process

• Stop the primary SAP HANA database system.

~$ sapcontrol -nr 0 -function StopSystem HDB

• Start the takeover on the secondary SAP HANA database system.

~$ hdbnsutl -sr_takeover

• Register the former primary to become the new secondary.

~$ hdbnsutil -sr_register --name=DRSITE01 --remoteHost=saprep02 --remoteInstance=00 --replicationMode=async --operationMode=logreplay

• Start the new secondary SAP HANA database system.

~$ sapcontrol -nr 0 -function StartSystem HDB

Post Migrate

• Wait some time till SAPHanaSR-showAttr shows both SAP HANA database systems to be up again (field roles must start with the digit 4). The new secondary should have role "S" (for secondary).
• Tell the cluster to forget about the former master-slave roles and to re-monitor the failed master. The command could be submitted on any cluster node as user root.

saprep01:~ # crm resource refresh rsc_SAPHana_SLE_HDB00

• Set the master/slave resource to the status managed again. The command could be submitted on any cluster node as user root.

~# crm resouce maintenance rsc_SAPHana_SLE_HDB00 off