For many of us, Linux has been a welcome presence in data centers large and small. What do you do, though, when you start getting into more than you think you can handle ?
With the advent of ever-evolving technologies like SANs, virtualization and server consolidation, data centers are glowing with even more shimmering lights, and humming from the buzz of smaller form factor stand-alone servers, farms of virtual machine servers and rows of blade centers. The ease of using server templates, cloning and automated installations have definitely had a great impact on the number of servers you end up managing today. What may have been a 10:1 server to technician ratio several years ago has now changed, and enterprise sized server farms of several hundred machines are managed by just a handful of people. Simply put…if you can build it faster, better and cheaper, someone will take notice and expect more.
Along with the growing data center, you also have the rise of Linux as an enterprise level operating system. As more tech houses use this constantly maturing operating system, they run into issues like support, hardware compatibility and finding ways to get more bang for their buck using open source components in their existing infrastructures. Now, aside from finding better ways to manage your hardware, using tools, monitoring processes and other fun IT stuff, one of the biggest headaches IT has to face is keeping your machines up to date. Yes…we’re talking about patching.
When it comes to patching, Microsoft has the edge by far. Regardless of the number of patches Microsoft puts out every year, being the popular operating system that it is, it gets pretty good support from the industry when it comes to facilitating patch management. Aside from using Microsoft Update to patch your machines, there are plenty of third-party tools that support the Windows operating system. For Linux, on the other hand, you’ll only find a few third party tools. You can use the built-in update processes that the OS has to offer, but it can quite clumsy, especially if scheduling is required, or if there are package dependencies to consider. The few third party tools available can be rather limiting as well, since the majority only work with RedHat. You also need to deal with a vast number of machines in your server farm. How can you manage large scale patch deployments across thirty, sixty or even several hundred servers ?
In this article, I try to cover some of the basics of patching Linux using built-in mechanisms, what’s available in the third-party tool market and, some of the obstacles I’ve run into when trying to manage a small to large data center full of Linux servers.