We recently have moved to Azure Elastic Scale DB. So when we need to create another shard, our DBA splits an existing shard manually. So I have been given the task to automate that process so we can easily split an existing shard or merge multiple shards into one. I am learning and following Elastic database Split-Merge tool tutorial. Another tutorial here says under Requirements and Limitations section the following:
Currently, the shards need to exist and be registered in the shard map before a split-merge operation on these shards can be performed.
So I still have to create another shard manually and register it in shards map, for which i may have to use another tool. Similarly schema for all sharded tables and reference tables need to exist on the target shard prior to any split/merge/move operation.
So this functionality of split/merge is not automated yet or am i looking at wrong place?
I want all these operations to be automated. Can you refer me some complete software or tutorial for that?
Cbattlegear is correct -- the process is not yet integrated into a single tool. Currently the Split/Merge tool only handles the data movement operations. However there is a Powershell example of orchestrating the entire process -- including provisioning the target databases, adding it to the shard map, and invoking the split/merge operations. You can find it described in the testing section of the Split/Merge Tutorial.
Some customer simplify the process even further by maintaining an empty "golden copy" database containing all schema objects necessary for shards. Then when a new shard is needed, they perform a CREATE DATABASE AS COPY command to quickly create a new DB from the reference copy. This step can be scripted as well.
Finally, in addition to Powershell, there is a .Net API available for Split-Merge that allows the tool's operations to be integrated with other steps or APIs inside a C# executable. The API is included in the Split/Merge Nuget package and documented here.