Hawk Host Report – June 2011

Despite some recent blog posts, we did do a lot more in the month of May besides looking for Easter eggs in the software we use. Now don’t get me wrong, we do love to have some fun and, you guessed it, our idea of a good time is looking for those Easter eggs and other hidden gems. But through all that fun we managed to accomplish quite a bit internally, along with some nice customer facing changes.

Memory Upgrades

For the past 2 years now, we’ve run a standard setup on all shared and reseller hosting servers of dual Intel Xeon series processors (ranges from 5520s to x5675s), 4x300GB SAS drives in a RAID10 array, and 12GB of memory. This standardization is great and the quality of this hardware lets us provide a solid all around hosting experience. Over time we started to notice that our disk io was getting a little high, and our monitoring showed we were inching further away from our “comfort zone” as far as system performance goes. Our mindset has always been to fix small problems before they became big problems, because big problems almost always have a real world impact. So to prevent a potential big problem, we made the decision to upgrade all of our servers from 12GB of memory to 24GB. This upgrade has had an immediate impact in more than one way. We’ve essentially eliminated any io wait from our disks, the few instances of load spikes we had before the upgrade have disappeared, and the volume of monitoring / notification emails have dropped to virtually nothing. A handful of users have reported some significant performance gains in their applications but for the most part, it’s business as usual and that is just how we like to keep it.

Ratepoint

We signed up for a Ratepoint account to give you guys another venue to review us at a credible, trusted source. We currently don’t have a button / badge implemented on our site, but there’s an issue open with our webmasters for this and once we determine a clean way to do this we’ll have it right on the site. For now if you want to put in a review for us (which we’d really appreciate!) you can do so by clicking here.

Mercurial to Git

For the past few years we’ve been using Mercurial with Bitbucket for our code repositories, utilizing the repositories for things like internal development, staff projects and the like. This past month our team made the decision to move away from Mercurial and switch over to Git. As a result, we moved where our repositories are hosted over to GitHub, perhaps the most well known and trusted collaborative platform for hosting git repositories. While there’s a handful of reasons we moved over, it primarily came down to what our staff was most comfortable working with. Between our sysadmins and developers we all came to a mutual agreement that Git (and GitHub) were extremely well supported, they were all very familiar and comfortable with the syntax, and it also allows us a lot more control and flexibility with each of the repositories we’re hosting.

Hardware for the road

One thing that we’re all proud of here is how well we communicate and how effectively we work together, despite not having a centralized office (yet). As our growth has continued (and rapidly) we’ve really benefited from open and easy lines of communication between all staff. We use a lot of software and utilities to make sure we can get a hold of each other on a moments notice, and we also have a great amount of documentation and information on what any staff member is working on at any given time. All that in mind, the rapid growth we’ve seen in nearly every aspect of our business (customer base, servers, internal properties, staff, etc) has demanded that we re-evaluate our approach and see where we can improve on communication. One area we lacked was mobile availability, and so we made the decision to purchase laptops for our employees who at any time could be required to make a trip to our business office in Ontario, Canada or to our operations office in Buffalo, NY. Not only are these trips vital for keeping everyone in the loop and things moving, but it’s a great chance to get some face time with the co-workers we always talk to and almost never see 🙂

Miscellaneous

Our development team made some great progress on our internal system which has had a *major* impact on our workflow and the productivity of everybody on staff. We’re also heading into June with a readjusted work schedule so our live chat availability will be increased compared to what it’s been for the past few months.

All said and done May was a very productive month for us on all fronts and we couldn’t have asked for more from everyone involved. One last fun fact before I wrap this thing up: the number of orders we received last month shattered our previous record 🙂

This entry was posted in General and tagged , , , . Bookmark the permalink.

4 Responses to Hawk Host Report – June 2011

  1. Saqib says:

    Good going Guys. So the 24 GB RAM would definitely speed up things. And I would definitely write review on Ratepoint, as you people offer great service and support at such an affordable price. Keep going.

  2. vampireJ says:

    You should have hosted your own git repo

  3. vampireJ says:

    You do not intend to increase Pluto’s memory are you?

  4. Tony says:

    We will not be increasing the memory on the two remaining 5400 systems we have as we’ll be replacing all of the hardware eventually. We also stopped deploying new accounts on these machines quite a while ago so the user levels are not really close to what our newer machines have.

Leave a Reply

Your email address will not be published. Required fields are marked *