When we began building Contegix, we realized it was always about the applications.  No one buys a server (physical or virtual) to run a ping and get some ICMP packets.  The Internet has, and always will be, about the applications.  Therefore, we made the application infrastructure the core of our management and focus of our hosting services.

As we build out cloud templates, we want to extend this belief into the templates themselves.  It can not be about just the operating system.  Each template has been built “The Contegix Way”.  These are built using the components and deployment methodologies that have served our customers for years and continue to do so today.  Some of these target developers, some for system administrator implementors, and some target application users.  Below is our list, but we want your feedback.

For developers:

• Linux 64-bit with Apache HTTPd + Jetty
• Linux 64-bit with Apache HTTPd + Apache Tomcat 6
• Linux 64-bit with Apache HTTPd + Apache Tomcat 6 + Grails 1.2 (well, we can hope 1.2 will be done soon  )
• Linux 64-bit with Apache HTTPd + Phusion Pasenger + Ruby Enterprise
• Linux 64-bit with Apache HTTPd + Django

For system administrators / application implementors:

• Linux 64-bit with MySQL 5
• Linux 64-bit with PostgreSQL 8.4
• Linux 64-bit with MongoDB (probably… working on it!)

For application users:

• Linux 64-bit with MovableType
• Linux 64-bit with WordPress + MySQL 5

Yet, we want to hear – what do you want to see in a cloud template and ready to deploy in less than 5 minutes?  Please leave comments.

We have been chatting about our cloud computing platform (known internally as Project Zeus but to be official named later) a fair amount in the past few months.  We have talked about our features and it being truly designed for the enterprise – small and large.  We have not talked about how it was built.

Earlier this week, our development team at Relevance put up a blog post on their work on the web and mid tiers for the project.  This promises to be the first in a series on how to build a cloud computing platform.  At least, how Contegix built a cloud computing platform designed around our principles and requirements.

Read the post, and one will find a number of interesting jewels.  One thing might stick out as very interesting. While we are known for our Java hosting, our cloud was built using Ruby and Rails.  Relevance used JRuby in the mid-tier for  numerous reasons, including the VIJava library for VMware.  The web tier is pure Rails 2.3.2.  Since Contegix is one of the largest managed Rails hosting companies and has been for quite awhile (anyone miss the Apache+fcgi days?), this was a natural infrastructure for us to support.

More information coming about our cloud late next week.  Launch is quickly approaching!

The winners were not supposed to be announced until May 1st. Yet, you can not keep the publisher and the USPS from their missions…

We are very proud to announce that our customers and the readers of the Linux Journal have selected Contegix as the top linux-friendly web hosting company, ahead of some of the biggest names in hosting. The June 2009 issue of the Linux Journal features the results of the Annual Reader’s Choice awards, where the Linux community votes on top products, tools, and services in the market. Contegix had a higher percentage of votes than any other hosting company included in the survey. We are proud to share the “favorite” distinction with many other excellent Linux products, such as Ubuntu – the “favorite primary Linux distribution”, Apache – the “favorite web server,” and Python – the “favorite programming language.”

From the article:

Talk about a meteoric rise, Contegix went from one write-in vote in 2008 to champion of the Favorite Linux-Friendly Web Hosting Company category in 2009. Otherwise, four well-known names have remained among your perrenial favorites for years, namely Rackspace, GoDaddy.com, DreamHost and 1&1. Of the four, only Rackspace broke the 10% barrier this year for Honorable Mention, while the other three were just shy of the mark. GoDaddy.com was category winner last year, and DreamHost and 1&1 were the two Honorable Mentions.

Everyone here at Contegix would like to thank our customers, our partners, and the members of the Linux community who voted for us. This award would not have been possible without your support. You trust us everyday to manage your infrastructures. We are grateful for that privilege and the award you bestowed upon us. You are the reason we do what we do.
Finally, we would like to thank the Linux community in general for continuing to develop the technologies that we know and love and that make us passionate about being the best Linux hosting company we can be.

(Picture of Phil Jacobs from StyleFeeder who sent us our first congrats tonight.)

In the past few months, nginx (pronounced “Engine X”) has become The Little Engine That Could. This is most evident in Rails deployments and in Zimbra 5, where it replaced perdition for IMAP/POP3 proxying. For Rails, it is typically replacing Apache 2.2 proxy_load_balancer as a front-end to Mongrel.

One of our engineers, Joe Williams, decided to put both system to the test with a Battle Royale. Check out the results.

Beginning in September, we began to quietly upgrade our backup system. The previous system worked well and never lost customer data. It was the perfect system when we purchased it in December 2005. It suffered a few problems, including slow startup restore times and a sad dependence on filesystem inodes for the indexing system. This led to scaling issues that would have drastically impacted us by mid-2008.

After a lengthy investigation into option, we decided to use NetVault from BakBone Software. The decision was made due to fulfilling specific criteria – native Linux client and server software, support for MySQL, PostgreSQL, and Oracle hot backups, Disk-to-Disk-to-Tape (D2D2T) capabilities, and the ability to implement quickly, very quickly. The last one was key to minimizing our costs for the implementation as we knew we would need to maintain the old system for a minimum of 60 days after the last server was migrated. This also led to RedHat and Bakbone producing a case study.

Feel free to read it here.

In an ongoing effort to improve service to customers, Contegix will be making some upgrade infrastructures to our core DNS services. Effective October 20, 2007, Contegix will be making changes to cache value, called TTL (http://en.wikipedia.org/wiki/Time_to_live#Time_to_live_of_DNS_records), of all DNS records to one hour (3600 seconds). In the past, most TTL values were set at 24 hours except during migration or changes where this value was lowered in order to minimize downtime. This was the Contegix default.


Hopefully, these FAQ below will help answer any questions:

Q: Why is Contegix making this change?

A: This change is being done in order to help facilitate quicker changes to customer DNS records. Customers previously had to wait up to 24 hours to see the change replicate among all DNS servers throughout the Internet. With the change, this time has been shortened to 1 hour.

Q: According to the Wikipedia link, lower TTLs cause heavier loads on the name servers. Is the infrastructure capable of handling this?

A: Absolutely. The core Contegix DNS infrastructure is capable of handling this load and much higher. Current name servers are sitting approximately 99% idle with an average response time of 5-8 milliseconds. This level is being held even with TTLs at 600 seconds or lower for over 4,000 DNS records. In addition, the change for reverse DNS (IP to hostname lookup) has been using the new value for over 6 months with no noticeable load difference.

Q: Will Contegix still allow custom TTL values – lower and higher?

A: Absolutely. Contegix will continue to offer customers the option of having lower and higher TTL values. Contegix remains one of the few providers who allow TTL with values below one hour. Please file a support request if you wish to have a different value set for an entry after the changeover.

Q: What action do customers need to take?A: Unless the customer has specific TTL requirements, no action is needed upon the part of the customer.

Q: What if I do not host my DNS with Contegix?

A: No changes are required on your side.

Q: I like what I hear about Contegix’s redundant DNS infrastructure. Can I move my domain(s) there?

A: Contegix provides DNS hosting as part of our services. You are welcome to move your DNS hosting to us. Contact support at your convenience and the engineers will guide you through the process. Please do not change your DNS until you have contacted support and planned a migration strategy.

As always, thank you for your continued business. We look forward to helping you build the best infrastructure available. As promised, this is one additional announcement on our continued improvements to drive more customer value.

Our single most important commitment to our customers is to provide the absolute best support possible. Part of our support commitment is continuously pushing the response from reactive to pro-active. We want to be able to address an infrastructure problem before our customers realize it is a problem. When it is an application issue, we want to be able to provide you every piece of data available to help you resolve the application level issue. A key component to this goal is our monitoring system. With this goal in mind, I am please to announce that we are making a significant improvement to this core infrastructure.

We have selected Hyperic (http://hyperic.com) as the foundation of our new monitoring infrastructure. Specifically, we are deploying the Hyperic HQ Enterprise Edition (http://hyperic.com/products/hq_for_ent.html). Hyperic HQ drastically increases our monitoring capabilities by focusing on the health of the server and the applications in your infrastructure. Purpose-built for web infrastructure, and architected to consider all layers of infrastructure including hardware, middleware, virtualization and applications, Hyperic HQ delivers system monitoring, trending, and analysis.

So what does this mean for you – our customer? Let’s look past the server health and examine the application infrastructures:

• For our customers utilizing Apache, we will now be monitoring the health of Apache, including its availability and number of available Workers/Servers.
• For our customers utilizing Java, we will now be monitoring the health of the JVM via JMX. These statistics will be collected to determine the health of your application.
• For our customers utilizing Rails, we will now be monitoring the health of individual Mongrel instances.
• For our customers utilizing Zimbra, we will now be monitoring the individual components, including availability, speed of IMAP connections, and LDAP connections.

All of this means faster support, faster resolution, and great visibility. And that’s just the beginning…

We have decided to raise the bar on the managed hosting industry and deliver even more value through our trademark Beyond Managed Hosting. For our managed customers, the monitoring will continue to be included in your services. This drives more value from the management services we provide with more visibility into your infrastructure. Beginning this weekend, our support engineers will begin making the necessary changes to customer firewalls to accommodate the Hyperic agent.

The rollout of the agent will begin in approximately two week with completion by March 15, 2008. After the rollout is complete, direct customer access to the system will begin rollout with completion by end of March 2008.

If you are a colocation customer, feel free to contact your account executive or sales@contegix.com to find out how you can take advantage of this monitoring solution at a reasonable per-device cost.