Following my previous post, I’m explaining how to use Azure Blob Storage for OpenShift internal docker registry storage.Read more
Here is what I did for creating my demo cluster for OpenShift Container Platform 3.5 on Azure.
- Three masters with etcd (same hosts). These masters are also infra nodes.
- One node (We can add more nodes later if we need)
- One Azure Load Balancer with two public IP (master endpoint and router endpoint) We can’t separate these endpoints into two ALB because Azure doesn’t allow creating two or more ALBs with one backend pools.
While I'm reading the document, I found difficulty in resource file naming.
To summarize first:
- "name" property of the project.json uses as the assembly name of the .NET Core project. The top level folder name is used if this property is not specified.
- We must abbreviate the assembly name from the resource class FQDN if the FQDN starts with the assembly name.
- We must use full FQDN if the FQDN doesn't start with the assembly name.
This blog describes two things.
- How to store session data out of the ASP.NET Core process
- How to separate the life cycle of session data from that of ASP.NET Core process and host
In the background, ASP.NET Core app can run as a container (docker) architecture. When ASP.NET Core app runs as a docker, it's easy to scale up and down. In this situation, we would like to store session data out of Docker containers and separate lifecycle. However, you may think if we can store session data out of the ASP.NET Core process, we can separate the lifecycle of session data from that of ASP.NET Core process and host at the same time. However, this is not necessarily so at present ASP.NET Core. I explain it based on the Microsoft's document.Read more