Posts

Showing posts from May, 2022

Upload your image on Azure Container Registry

Image
In my previous article , we went through the process of uploading our own image to Docker Hub. Even though it's perfectly fine to use Docker Hub, some organizations prefer using their own registries, either due to additional functionality that they might offer or due to policies and regulations. Since Azure offers a container registry resource that would fit most use cases, I thought we could try to upload some images and at least use it for test and dev purposes!  First things first, the tools we're going to need. To upload an image to an ACR we need: 1. An Azure Container Registry 2. Docker Desktop 3. Azure CLI Starting from the top of the list, we need the ACR resource to which we're going to upload the images. Clone my Github bicep repository and switch to the  ContainerRegistry-ImageUpload-001/110-Bicep/110-AzureContainerRegistry/ folder. Update the parameters in the deploy script and submit the deployment. The output should be similar to the below: If everything goe

High Performance K8s Storage with Rook & Ceph

Image
Modern applications are usually developed with cloud-native principles in mind, there are however some that may have particular requirements in terms of storage. When developing for containers, the need for a ReadWriteMany storage may arise, which may turn into a problem when cloud services fail to match the requirements, especially the ones related to performance. One of the solutions to this problem is the Rook - Ceph combination. Ceph is an open-source software-defined storage platform that implements object storage on a single distributed computer cluster and provides interfaces for object-, block- and file-level storage. Rook, on the other hand, helps perform all the administrative tasks such as deployment, configuration, provisioning, scaling, and more. A multi-zone AKS cluster is the perfect home for Rook and Ceph. The cluster is spread across three data centers and so is the data handled by Ceph. This increases the SLAs of the services running on the cluster and at the same tim