[ClusterLabs] SAP HANA resource start problem

Andrei Borzenkov arvidjaar at gmail.com
Sun May 14 02:26:09 EDT 2017

12.05.2017 13:30, Muhammad Sharfuddin пишет:
> is there a bug in SAP HANA resource ? crm_mon shows that cluster started
> the resource and keep the HANA resource in slave state, while in actual
> cluster doesn't start the resources, we found following events in the logs:

SAP HANA agent implements decision logic when and whether it can start
database internally; this logic cannot be mapped to pacemaker states
(pacemaker simply does not offer necessary abstractions) so you can
forget about using pacemaker to monitor HANA states. Use supplied script
show_SAPHanaSR_attributes to see states. start/stop on pacemaker level
simply initiate periodic actions; database is started by monitor action
when it becomes possible.
> 2017-05-12T01:02:02.650676+05:00 saphdbtst1
> SAPHana(rsc_SAPHana_TST_HDB00)[27645]: INFO: LPA: landcape: DOWN, LPA:
> wait ==> keep waiting
> 2017-05-12T01:02:02.659956+05:00 saphdbtst1
> SAPHana(rsc_SAPHana_TST_HDB00)[27645]: INFO: RA: landcape: DOWN, LPA:
> wait ==> keep waiting

Here is what you need to investigate.

More information about the Users mailing list