Kernel settings, schedulers, and solid state drives

Curtis — Wed, 08 May 2013 07:27:38 PDT

I’ve been testing various commercial solid state drives for a project that involves stateless, mostly Windows-based, virtual machines (vms). One thing about Windows vms is that they can use tens of thousands of IOPS while booting, which is part of why we are investigating solid state drives. Because SSDs are so fast, there are few kernel settings, and even scheduler changes, that can be used to improve performance. I find that most kernel parameters have great default settings, but are really set for generic use. Once we start getting over 100,000 IOPS, it might make sense to make a few changes to the kernel parameters, or even change the scheduler. At least, these changes and their effect on performance are what I’d like to investigate. Note that there are many variables involved here, including the page sizes the SSDs are using, or the chunk size of the mdadm device, among others. Some variables are just completely unknown. Curtis Collicutt tests various kernel settings on the performance of solid state drives. read more

[[ This is a content summary only. Visit my website for full links, other content, and more! ]]

IOPS in a windows boot storm

Curtis — Wed, 17 Apr 2013 10:45:06 PDT

What we are going to do We are going to boot 30 Windows 7 virtual machines — each with one virtual CPU (vCPU) and two gigabytes of memory — in two “boot storm” tests: one where we boot the virtual machines (VMs), 240 seconds apart, and the second where we boot 20 seconds apart, and watch the IOPS use while that is happening. Curtis Collicutt describes doing two "boot storm" tests with 30 Windows 7 virtual machines, booting at 240 then 20 seconds apart, and watching the resulting IOPS use... read more

[[ This is a content summary only. Visit my website for full links, other content, and more! ]]

The Rise of the Meta Operating System

Joe Topjian — Tue, 09 Apr 2013 08:06:57 PDT

Yo Dawg, I Herd You Like Servers, So We Put a Server in Yo Server so You can Serve More Servers Like an Xzibit Meme, server virtualization is full of recursion. Let’s look at the components and see how. First, there’s the virtual machine “host”. The host is usually a physical computer — although it can also be virtualized — that runs something called a “hypervisor”. A hypervisor is a software component that manages virtual machines. Then there are virtual machine “guests”. These are the virtual machines that run inside the host. The following image from RedHat depicts this structure: Server virtualization is full of recursion. Joe Topjian looks at the components to explain how this works... read more

[[ This is a content summary only. Visit my website for full links, other content, and more! ]]

The best development tool is...Internet Relay Chat?

Curtis — Thu, 04 Apr 2013 14:21:35 PDT

(complex use of the weechat IRC client) Cutting edge questions The more I work with cutting edge technology — which, for me, means software that is still being heavily developed and doesn’t have good documentation yet, or software that isn’t commonly used — the more I find Internet Relay Chat (IRC) to be helpful. In terms of organizational culture, Cybera likes to investigate new technology early in the adoption curve. That means asking questions! Many, many questions. And where’s the best place to ask questions about new, fast developing, or underused technology? I think it’s IRC. A few examples of software I use almost daily are: "The more I work with cutting edge technology," writes Curtis Collicutt, "the more I find Internet Relay Chat to be helpful...." read more

[[ This is a content summary only. Visit my website for full links, other content, and more! ]]

Covering up complexity

Joe Topjian — Tue, 02 Apr 2013 11:10:28 PDT

Over half of the jobs I’ve held during my IT career have been with public service providers — whether providing internet or web hosting services, email hosting, or now, cloud services. I have always found these jobs the most fun and rewarding due to the creativity that can be applied. For example, as a junior systems administrator for a small ISP, rather than sitting on the phone with customer after customer and manually walking them through setting up an email account, I was able to build a small web application for them instead. It allowed customers to create and manage their emails, as well as download dynamically generated scripts that would configure their desktop for those addresses. What happens when a complex system becomes so stable that the intricate inner workings can safely be ignored by the operator? Usually, the system is then marketed as a package.... read more

[[ This is a content summary only. Visit my website for full links, other content, and more! ]]

Welcome to the (web) jungle

Meagan — Thu, 28 Mar 2013 07:44:36 PDT

Web management is not for the innocent or naive. I am slowly learning this. There are many con artists with tricks up their sleeves, and you need to stay on your toes to avoid falling into their traps. Personally, I have always found it easy to ignore the poorly-worded messages from faraway lands offering to improve my SEO: “Do you want to see your website in Top 10 positions in Google or other major search engine? “If yes, please let me know your website URL and keywords name that you want to optimize so that our team could supply you with our best price to optimize the website….” I equate messages like these to business offers from exiled princesses: fit for the bin. But sometimes, there are actions that catch me off guard, and I have to turn to my tech team for advice. Meagan Hampel navigates the world of web management, learning to avoid sales tricks and online hoaxes... read more

[[ This is a content summary only. Visit my website for full links, other content, and more! ]]

Finding love in monitoring

luketymowski — Mon, 18 Mar 2013 15:53:04 PDT

Monitoring, long an operations backwater, began to get some attention in early 2011, but for all the wrong reasons. None of the tools had made any progress since the 1990s. “Monitoring sucks!” was a common complaint on Twitter (the #monitoringsucks hashtag began to trend at one point). When all your servers are physical machines, and you’re watching them from the same network, monitoring is not too painful. You may realize that there are better ways to do it, but that would require a lot of work, and in operations, your job is to maintain services and the servers necessary to provide them. Building ambitious new tools is usually not part of your mandate. Luke Tymowski outlines the many monitoring tools now available for network operators... read more

[[ This is a content summary only. Visit my website for full links, other content, and more! ]]

Internet routing and data security: the latest from the industry

Alvaro — Tue, 12 Mar 2013 15:48:59 PDT

The North American Network Operators’ Group (NANOG) conference has been running three times a year since 1994. I attended my first NANOG conference in February of last year, and found the content, and especially the people, to be awesome (being hosted in Orlando certainly helped). I was able to meet people who pretty much run the core of the Internet: Tier 1 ISPs, Domain Name Registrars, vendors and developers. The mood was great, and although most attendees are technically competitors, they all know they are there to make the Internet work better. Think you know everything about domain name systems? There's still some important facts to learn, as Alvaro Pereira discovered at last month's NANOG... read more

[[ This is a content summary only. Visit my website for full links, other content, and more! ]]

Eight people, five days, one OpenStack book

Joe Topjian — Thu, 07 Mar 2013 00:27:35 PST

This past week, I had the opportunity to participate in a Book Sprint. A marathon of sorts, a Book Sprint is where a group of people dedicate an entire week to writing a book from start to finish. The goal of this Sprint was to write an operations manual for OpenStack. While documentation exists for installing and developing code for OpenStack, there has never been a definitive source of information for people in operations. For example, topics like designing a cloud, monitoring a cloud, and troubleshooting issues in a cloud have been scattered or kept internal to individual organizations. To resolve this, five OpenStack operators and three documentation experts met in Austin, Texas to set out to write the official OpenStack Operators Guide. Joe Topjian describes the Book Sprint to create the first official OpenStack Operators Guide... read more

[[ This is a content summary only. Visit my website for full links, other content, and more! ]]

Chocolatey: A package manager for Windows

Curtis — Tue, 26 Feb 2013 08:38:39 PST

I have worked my entire career as a Unix/Linux systems administrator, so it was a challenge when I started working at Cybera and was placed on the Virtual Computing Lab (VCL) project. VCL, though running on top of OpenStack, is currently providing students with access to 100% Windows 7 based virtual machine instances. Some day, when a class calls for it, I’m sure we’ll provide students with access to Linux images, but for now it’s completely Windows. This means that I’ve had to learn a lot about the Microsoft ecosystem, and considering I haven’t used Microsoft operating systems since Windows NT, there is a lot of information to cover. Getting to know Chocolatey, an apt-get or yum-like software package manager --- for Windows! read more

[[ This is a content summary only. Visit my website for full links, other content, and more! ]]