Chris Kernaghan knows SAP Vora (Vora). To be honest, he’s a little obsessed with this product. It’s this passion, combined with his experience of multiple Vora deployments, that made Chris the ideal person to go to when we wanted someone to share their experiences of Vora to help you in your journey...
It does not feel like I have been working with Vora for 18 months; it feels like a few weeks since I built my first Ambari Cluster and deployed what would become Vora. I was blown away to say the least, and instinctively understood why this product was going to be so exciting to work with.
Fast forward to today: we are already on Vora 1.4 and have seen some amazing features added, with additional ones on the roadmap. This product is continuing to move forward at an impressive pace. Here are my observations and learnings from working with a product that is advancing so quickly. In no particular order these are things I have learnt when deploying Vora since it was released.
Pay attention to the version numbers
SAP HANA Vora sits on a multi-layered stack of products; SAP has partnered with several companies who provide these products as a certified stack (Hortonworks, Cloudera, MapR). Using the SAP PAM we can see the supported configurations for all the versions of SAP HANA Vora.
Where I have often seen customers having issues is when they already have a Hadoop cluster installed and are trying to put Vora on top of it. Working with these installations is interesting because often you have to double check the version of the Hadoop components, like Spark, to verify supportability. SAP do not support every patch for a component, for example Spark 1.6.3. This is predominantly because they can’t test everything, so they pick their supported patches.
There is a hierarchy of information correctness
With every product there is a set of documentation and supporting materials. With products that are developing quickly it is vital to recognise an information hierarchy. This is vital in helping you to understand which instruction to follow when deploying or troubleshooting an issue.
For Vora the hierarchy is - SAP Master Guide, SAP Help (HTML), SAP Help (PDF), SAP Notes.
Don’t forget there are some great videos on the SAP YouTube channel, but make sure you are watching for the right version!
Patches, patches, patches
The SAP HANA Vora development team regularly provide patches for SAP HANA Vora on the Service Marketplace. These patches are a mixture of fixes and new tweaks to features. This is a very useful process, although it does pose a few challenges in environments where change and release cycles are long.
Contacting the product team
The Vora product team are engaging directly with users through a number of channels.They answer questions on SCN, StackOverflow in addition to Twitter. They also have a message queue on Service Marketplace.
Vora sits upon technology which is frankly alien to many within the SAP technical arena. Hadoop is a challenging and different animal but there is a lot of information out there for people to leverage. Personally I have used several resources, Stack Overflow, YouTube and Hortonworks Forums. Likewise the Vora development team have a great FAQ on troubleshooting.
SAP HANA has numerous data types.It allows for an exceptionally granular control of the amount of storage of data within the database. Things can get a little tricky, however, when moving data from HANA into Vora because the data types are not exactly aligned because of its basis in Spark. For example; the data type SmallDecimal does not exist in Vora. Therefore there needs to be a translation, but this is quite normal for data migration and mapping.
Recently I have given a presentation to the SAP User Group Big Data and Analytics members on the importance of a Big Data Lab to learn and test ideas and as I write this my whiteboard at home has a new set of tasks to upgrade my own data lab to include Vora 1.4. The new features look awesome, especially the improved installer and the Vora manager updates.