One of the biggest gripes many people have with in memory databases is the following. What if one of the SAP HANA blades experience catastrophic failure and are no longer usable. Since that data is stored in memory, what happens to that data? Have we lost it forever? Will we need to recover from a hard disk backup? SAP has figured out a solution to this problem.
Let’s say we have a SAP HANA set up consisting of eight active servers: HANA01 – Standby Server salon hana (becomes new active Index Server for failed HANA05) HANA02 – Master Server – Index Server HANA03 – Index Server HANA04 – Index Server HANA05 – Index Server HANA06 – Index Server HANA07 – Index Server HANA08 – Standby Server
For testing purposes lets say we remove HANA05 from the mix. As soon as the SAP HANA system senses that HANA05 has become disconnected from the cluster it will immediately be transferred to a standby salon hana server. HANA01 will now transform from a Standby Server to a full blown Index Server. At this point in time it will read off of the hard disk backup that HANA05 was writing to. This allows for a seamless recovery that normally takes around five minutes to run.
It’s a best practice with SAP HANA to have a disaster recover plan at your place of business that should incorporate using a remote disk that is offsite from the data center. If all backups are destroyed, the data will need to be reindexed. salon hana Click Here For SAP BW Training Course star rating Click Here For SAP Training Course star rating
No comments:
Post a Comment