Earlier I posted asking What Happened to Vertebra?, the sysadmin cloud management tool that EngineYard had promised for ‘end of summer’. Pleased to see that the EY folks soft released an early BETA of the code today. You can see more at their github repo.
To be fair, this is the very first external release of this tool, so you shouldn’t be setting expectations too high nor should you be playing with it unless you’re an early adopter and comfortable with dealing with complex sysadmin tools. Also, most of the core is written in Erlang, which is still not a widely adopted language.
Pretty excited to take a close look. There is a lot of promise here.
Update: Here is the
official Vertebra home page, which gives more of a high level picture of the tool.
December 19th, 2008
Earlier this year during the 2008 Velocity Conference EngineYard made some fairly big waves (Google search) pre-announcing their open source cloud management tool, Vertebra.
This was of particular interest to me because of the CloudScale Project. We spent a considerable amount of time building a sysadmin messaging bus, but were never 100% happy with it. Vertebra looked like just the ticket, it would be open, and hopefully widely embraced, perhaps making our lives easier.
At the time, the EngineYard (EY) folks promised ‘right after summer’. With December literally right around the corner, you’ve got to wonder whether this vaporware will ever really show up. Given the recent layoff of the Rubinius team from EngineYard it seems logical to conclude that their management realizes that too many over-funded R&D projects aren’t necessarily going to be in the companies long term best interests.
Merb made it out the gate thankfully, but will Vertebra? There hasn’t been a single public release since it was pre-announced almost 6 months ago so I’m going to hazard a guess that we’ll either never see it or at least not for a long time. Maybe Velocity Conference 2009??
Come on, EngineYard, just get it out there, warts and all. Either it will gather some more support and live or it won’t.
Update: EY just released Vertebra today (2008/12/19). Find out more
here.
November 30th, 2008
Looks like Don MacAskill discovered the joys of MySQL on ZFS running on OpenSolaris.
This is great stuff. When we release Scaleen later this year as part of the CloudScale Project, it’s going to make managing ZFS+EBS on EC2 trivial and fun.
October 11th, 2008
It’s been a while since my last post and a lot has happened for both neoTactics and CloudScale. I’m quite excited to make several announcements about the CloudScale Project here. Although it’s not the announcement I had hoped to make, that we had received venture funding, I am pleased about our new direction.
Background
neoTactics has been incubating CloudScale as part of our cloud-onboarding engagements with several clients. It was intended as a next generation cloud management platform, similar to RightScale or Elastra, but architected and developed by experienced Internet and web operations professionals. We designed it to solve the generic problem of managing groups of servers that form web applications on any cloud.
CloudScale reached a final ALPHA very recently, but for a variety of reasons we have decided to open source parts of the project and perhaps the entire solution in the future. We are currently evaluating our options, but you will see bits trickling into the open source community soon.
While CloudScale may not emerge as a commercial entity in the near term we feel that releasing parts of it will be beneficial for everyone.
The First Bits
The very first bit we’re releasing, in conjunction with David Schmitt of das Z, are some extensions to Puppet to add support for configuring collectd.
We commissioned this work from David to help accelerate our progress on CloudScale. If you use Puppet, also be sure to check out his extensive list of other Puppet modules.
David’s announcement can be found on the Puppet users mailing list.
Monitoring & Configuration Management
The release of the collectd extensions for Puppet is very important. Historically, monitoring and gathering performance data from servers is a ‘bolt-on’ activity where the implementation is added after the fact. This is a huge problem on clouds. If you want to turn on 10 servers for a day, then spending a half day configuring performance monitoring is a huge amount of overhead.
collectd is one of the few monitoring and instrumentation tools available that inherently works in a distributed, cloud-friendly fashion. By making it trivial to configure collectd with Puppet, authors of Puppet modules can now create not only a recipe for installing applications like Apache and MySQL but also include the recipe for monitoring those same applications! Puppet is, of course, the only rational tool for doing any kind of significant cloud configuration management and provisioning.
Making these two tools work closely together will make it much easier to develop robust cloud and datacenter infrastructure.
More Soon
Stand by. There will be a bunch of new blog posts and releases coming soon. Now that we’re not spending so much time attempting to raise money we should have lots more time to post and tidy up the bits of code we want to release.
Good times!
October 6th, 2008