HANA High Availability System Vs Storage Vs VIP failover

[Former Member]

Dear Experts,

Hope your all doing great. I would like to seek your expertise on HANA high availability best practice. We have been deciding to use TDI for BW on HANA. The next big question for us is how make it available atleast 99.99%.

I was going through multiple documents, SDN forums, etc..but would like to see how the experts are performing in real time.

My view -

Virtual IP failover is a common HA practice which have been used to failover CI / DB hosts depends on failure/maintenance. In this case, both nodes can be used to run app servers.

System replication - HANA based required secondary standby node, which doesn't accept user requests, but replicate the database from primary using logs after initial data snapshot either synchronous or Asynchronous. (Can be used as HA or DR - if servers are between different data centers).

Storage replication - HANA based required secondary standby node, which replicates SAN for HA/DR.

Could you please provide your expertise method you followed for HANA HA and what are the pros, cons and challenges that you have faced or facing.

Thanks

Yoga

[Former Member]

Hi Yoga,

Did you got the answers by this time ?

All the three methods i.e., has it own pros and cons with different RTO and RPO. For example if it is a shared storage and failover between two nodes i.e., active-passive we will have problems if there is a issue or a problem with the storage itself. 

SAP currently recommending to go for a system replication and with a full sync mode through which we can achieve almost zero RPO .  Added to that SUSE offering a automation of system replication between the nodes.

Hope this answers your query.

Thanks & Regards,

DVRK