Real Server Lifecycle Finally Arrives
A deep system model enables lots of things (fast provisioning
, incremental updates, and compliance to name a few). But my favoriteand the most profoundis system lifecycle management that actually works!
Developers are accustomed to straightforward application lifecycle management thanks to version control. A typical day-in-the-life for developers these days includes activities like:
Promoting a build to testing
Creating a version branch against a prior release to fix a bug
Marking a successfully-tested version as production ready
Release management for developers used to be total chaos, but now its manageable thanks to tools that control the lifecycle.
So why isnt that lifecycle savviness common in operations? Its because systems (and multi-system business services) arent modeled. When you promote a build to testing, you need some concrete object to promote.
Whats that object when youre talking about a system build? When you mark a good build as production ready, whats that build when youre talking about a complete 3-tier app deployment?
There are two answers to this question floating around right now.
One is the new DevOps approach: Script up the entire system deployment, then put the scripts under version control. I think thats a fine approach for systems where (1) the app is sufficiently simple to script up, and (2) the staff is sufficiently savvy at scripting.
The second is rPaths favorite: Use a tool that can declaratively model an entire system blueprint. Now you have that something to move back and forth through the system lifecycle. This approach works for messy unmanageable appsand doesnt require a staff of programmers to implement.
Heres how a lifecycle-enabled workflow looks for deploying a JBoss update:
Alice, an application engineer, takes the current version of the JBoss app server blueprint on the Production branch, makes the change, and checks it in as a new version on the Development branch.
When satisfied with the change, Alice promotes the version to the Testing branch. Promotion is an explicit feature in rPath.
Bob, a QA engineer, now has access to the new JBoss stack, and runs his tests. When ready, he promotes the stack to Production.
Now rPath shows that 56 deployed JBoss application servers are out of date since a new version of their blueprint is available. Carl, the release manager, schedules a change window, and Diane, a system administrator, orders rPath to update the systems to the new blueprint.
rPath analyzes each of the 56 systems and makes the minimal, incremental change necessary to bring each system up to date. Depending on the scope of Alices original change, it might replace a single file or an entire directory tree.
The most compelling benefit of this approach to lifecycle management is the complete assurance that Diane deploys only tested changes, and Bob is testing the right code. With traditional manual handoffs, its impossible to prevent configuration drift at each handoff.
Thats our vision of a real, developer-style lifecycle that any enterprise operations group can consume. I love seeing the aha moment when customers and prospects discover this.
by: rPath
How The Smtp Servers Work? Reasons Why Unlimited Web Hosting Is The Best Hosting Plan For You? Reasons Why You Should Choose A Good Web Hosting Company KBM2USB Facilitates Managing of Headless Servers What Is A Web Server? 3 Traits That Make A Fantastic Server Hosting Company How Multiple Ip Hosting Can Help Serp Rankings Seo Hosting Reviews How To Read Them Importance of Selecting Geographical Base Dedicated Servers Different Types Of Vps Server Hosting What are the Benefits of Virtual private server Web hosting? All About Domain Name Servers Ways to Serve: Collocated Servers and Dedicated Servers