Greetings from the Machine Room! We’re excited to share what’s been going on here at gridscale over the past few weeks.
Intended to be the central knowledge base for all features and functionalities of our products here at gridscale, our recently announced gridscale Product Documentation saw some updates.
The GSK Managed Kubernetes section now details how you can work with Kubernetes container logs. We’ve also added FAQs across several sections, and will be updating them and other sections as well over the coming weeks.
The Partner API now allows you to manage favourites for Contracts and Users, significantly reducing the time and effort required to search and manage frequently visited contracts and users.
We’ve also updated our API Documentation detailing password requirements for OS templates.
GSK – Managed Kubernetes
We’re happy to announce that GSK now supports versions 1.20 and 1.21! To simplify the upgrade process, we have shortened the total upgrade time for all clusters and versions.
We’ve also added custom SSL certificates via service annotations for Loadbalancers, available for clusters running the 1.19 patch version or newer. In addition to this, the time required to generate your kubeconfigs has also been significantly reduced.
If you wish to read more about these new changes, check out the gridscale Product Documentation.
Deprecated Resource Handling
GSK allows you to monitor underlying resources such as servers and storages that are required to run the clusters. These displayed resources for GSK clusters running versions 1.19 or later will be made read-only and thus imply that they are henceforth uneditable. This means that server configurations such as the number of cores and memory cannot be modified via the server itself, and neither can the firewall rules be updated.
The primary reason for this restriction is to ensure that stability is not compromised. In the past, it was possible to add a firewall rule which could potentially lead to cluster resource provisioning failures, or changes made to servers would be rolled back by Kubernetes during the autorecovery or update process.
This stability improvement, along with the recent horizontal and vertical scaling release, makes working with GSK not only a lot more stable but closer to native Kubernetes as well. We are also working on further improvements such as natively solving RWX volumes without having to manage your own server.
Preventing Potential Pod Failures
We’d also like to highlight an issue that’s been natively present within Kubernetes since 2018. Kubernetes mounts secrets such as service accounts within `/run/secrets/kubernetes.io/serviceaccount`. Since secrets are read-only, pods could potentially crash if they were mounted in a location such as `/run/secrets`. In order to prevent this from happening, we’d recommend using the workaround mentioned within the issue itself.
We’ve added ArchLinux 2021.07.01, Rocky Linux, NetBSD v9.2, OpenBSD v6.9 and CentOS v8.4 images. We’ve also improved the bootup performance for Debian-based templates.
We’ve refreshed the invoice layout, improving readability and making it easier to have an overview of essential billing information.
We’ve included current usages within Usage Dashboard. You can now see running usage per hour of all currently active resources associated with your account.
If you are a user without billing activated, then you no longer have the possibility to request a Virtual Location. Navigate to the Account Settings and add your billing details if you wish to take advantage of this dedicated hardware offering.
Within the Partner Panel, we’ve significantly improved the Dashboard to graphically display a better understanding of your cloud in terms of key financial metrics including Sales, Revenue as well as Costs incurred.
In case a server that you’re running ends up becoming inactive, it will no longer have VNC access within the Expert Panel.
PaaS – gridSQL
As part of the optimisation of our Platform Services, we have improved the reliability of Microsoft SQL Server database migrations.
We’ve optimised the platform’s default Q35 profile to adeptly run asynchronous and parallel workloads by enabling I/O threads.
If you’re interested in reading the previous post in this series, click here.
To stay up to date with recent news, you can subscribe to our RSS feeds:
Alan James | Team Product
Alan draws from a strong background in Product Management, Strategy, Operations and techno-functional FinTech consulting. While pursuing his Bachelor's in Computer Engineering from one of India’s leading technical universities, Alan was the lead for Google Developers Group.