What’s new in High Availability and Disaster Recovery with SAP HANA-SPS10

Posted by Prasad Illapani on

More by this author

With the release of SAP HANA-SPS10 in June-2015, SAP HANA customers/partners are leveraging the new In Memory Platform capabilities to run their mission critical applications. More information about the new innovations in the hana platform we have released is available in the SAP HANA-SPS10 blog written by Michael Eacrett. The focus of this blog will be towards the new enhancements and capabilities in SAP HANA-High Availability and Disaster Recovery areas of the platform.

1. Scale-Out, Persistence: Check for existence of vital SAP HANA files – Data and Log Volume  structures is now available. With every SAP HANA save point execution, a check of vital database files is executed where it is looking for the file/object existence displaying alerts early enough to start the activities and ease the impact of the accidental file deletion. The following objects that are checked in the process are:  Data volumes directory, Data volumes files, Log volume directory, Log volume segment files. With the new alerting capability in SPS10, HANA platform administrators can take early action in advance.

pic1

2. Scale-Out, High Availability (Host Auto-Failover):  Support for Dynamic Tiering with own failover group, which can be defined for each host where in the case of a failover, the nameserver tries to fail over to a host within the same group. In SPS10 with SAP HANA Dynamic Tiering, a special failover group for the group is added.
Failover group ‘default’
Failover group ‘extended_storage’
With this new setup option, the technology is extended to the different IO schema of Dynamic Tiering (shared everything) to offer Host Auto-Failover,  where Dynamic Tiering can now operate on NFS and SAN storage landscapes and no changes are required to the Storage Connector API.

Pic-2

3. Disaster Recovery, High Availability- SAP HANA System Replication:

A. Optimizations in Row store handling for faster take-over times: Extending the SPS09 feature keeping the Row Store in main memory on Secondary during take-over, with SPS10 the rebuild of the secondary indexes during reactivation of the Row Store now can be done in a decoupled way without waiting for finishing the creation process. It is done asynchronously in background where short-time missing secondary indexes can be coped with primary key accesses to involved tables until they are created.
Columnar Store (CS) is preloaded for possible take-over anyway, if the performance optimized version of System Replication is in use.

pic4

B. SAP HANA Cockpit: New Tile with Administration and Monitoring capabilities for SAP HANA System Replication is now available inside SAP HANA cockpit. Based on the type of the landscape (2 tier or 3 tier) the status of the services sums up the information underneath and presented as synchronization mode for the first replication in chain. The following options can be shown in the tile details:
Not configured (meaning system replication is not configured)
Active and in sync (green square)
All services are active but not yet in sync yet (yellow triangle)
Errors in Replication (red circle)

pic-5

Behind the main screen tile, there are lists of systems involved in the order of their relationships. The upper part sums up the chain of systems in case of Multi Tier system replication. The average write times are related to log writing where the values on ASYNC connects usually are smaller because of OS buffering.
Avg. Log Buffer size is the log information where SY stands for SYNC and AS stands for ASYNC.

Pic-6

The details of SAP HANA replication process per connection can also be displayed where vital data can be shown.  For ex: IO time’s history presented graphically with Shipped [ms] and Local [ms]. With a click on the diagram, columns related values and a link to alert details are presented to the user. More individual KPIs are also presented like: General Overview, Log Positions, Save points, Full Data Replica, Delta Data Replica and Backlog etc.

pic7

As part of our planned and future innovations, we are working  towards Operational security topic areas in a scale-out cluster landscapes and System Replication extensions – Hot Standby or Pure Log-based transfer, which acts as a foundation for Active/Active operations in a High Availability & Disaster Recovery scenarios.

 

 

VN:F [1.9.22_1171]
Average User Rating
Rating: 4.0/5 (4 votes cast)
What's new in High Availability and Disaster Recovery with SAP HANA-SPS10, 4.0 out of 5 based on 4 ratings

6553 Views