There are some universal rules for IT departments and one of those is that if downtime is involved, do it after hours. For decades, that has been the case for system updates because these updates take either the entire system or at least essential services offline to perform the update. With virtualization and clustering, however, that does not need to be the case and even more so when using a truly hyperconverged system.
Virtualization clustering is not new. It has been around for more than a decade, so this idea of no-downtime updates has been around that long as well, but it hasn’t always been easy. Because the hypervisor has been from a different vendor from the hardware vendor, hardware updates and hypervisor updates have needed to be performed separately. This complicates the update process because of unforeseen compatibility issues between patch levels in hardware and software. It also involved manually moving VMs between cluster nodes, making the process more tedious.
In a true hyperconverged system, however, the hypervisor and hardware are delivered by the same vendor in the form of an appliance. This approach allows the vendor to combine hardware and hypervisor updates into the same delivery package and make sure they are fully verified on the system before sending them out to your appliances. This eliminates the possibility of patch level incompatibility.
Also, with a fully hyperconverged system, the entire update task can be automated. VMs can be automatically moved from node to node so that each node of the cluster can be updated while VMs continue running happily, with no downtime, throughout the process. This is sometimes referred to as a “one-click” update and for the most part, it really is that simple. Not all hyperconverged vendors are the same, though and for many of these vendors, the “one-click” update is once for the hardware and once again for the software.
The Scale Computing HC3 system is truly hyperconverged and truly offers a “one-click” that updates the hypervisor, storage, and the hardware’s firmware in a single process. Updates are delivered directly to the management UI (unless you are in a dark site). Automation keeps the entire system online during the update, so there is no downtime of any VMs or workloads. The process can happily run during the daytime hours when people are working and users won’t even notice.
What this all means is that you, the IT administrator responsible for the HC3 system, should not need to schedule these updates for after hours. We have, however, been hearing that some of you are still scheduling these updates for after hours. Maybe you just aren’t ready to go home at the end of the day. Maybe this is your chance to get away from it all and sit in quiet solitude in the office sipping some coffee and checking up on social media. Maybe you are concerned about the performance of workloads during the update when they are moved to other nodes, and if so, we are happy to discuss those concerns based on your specific system.
If you are not motivated to work after hours, you can just do these updates during your normal working hours and we support that. We encourage you to check with one of our many ScaleCare experts if you have any questions about the process. Our goal is to make your IT administration as easy as it can be and we believe that includes freeing up your after hours to do something better than perform tedious system updates.
Guest Blogger Series: Our Move to Hyperconvergence - Part 3/5