I am a .NET guy with no knowledge about Linux, Apache, Java and Solr. But after research, I still think build a dedicated search server is the way to go.
What I want to achieve is: build a dedicated search server based on Solr and host on Azure Virtual Machine. I will use REST api to update the index and query. So for the server itself is set and forget (I hope).
Here is the plan after research, please advice.
Question 1: Get a Azure Virtual Machine, install Solr as this article advised. then where the persistent storage place (to put indexes), how to make sure it will not gone after the server reboot or redeploy.
Question 2: I want only my another instance in Azure can query, so is the Virtual Network come to play?
Question 3: Is this setup really maintenance free?
Question 1: The data will be persisted using a Cloud Drive (a page blob in Blob Storage) and this is persisted after a server reboot or redeploy.
Question 2: If you add more roles to that same hosted service/cloud service you don’t even need to bother with a Virtual Network, you can simply use internal endpoints. If you’re trying to access the hosted service/cloud service that hosts Solr from a different hosted service/cloud service or virtual machine, then you’ll need to use a Virtual Network.
Question 3: I’m not convinced that this setup is really maintenance free. Think of auto scaling your slaves, … Did you ever consider LucidWorks (which is a Windows Azure partner for Solr)?