If you are already running servers in a virtualised environment, then shifting into a cloud environment might seem like a relatively trivial step. However, there are some simple principles worth remembering to ensure the process happens with minimal hassle and that you can justify the costs involved.
Checklist picture from Shutterstock
This list is based on a presentation given by systems administrator Phil Ingram during the server administration mini-conference at Linux.conf.au in Canberra yesterday, which I'm attending as part of our World Of Servers coverage. Whatever cloud platform you choose (and whatever server OS you're starting from), they are issues that need to be kept at front of mind.
Make sure you factor in the cost of outages when comparing expenses. "Even with a 24 by 7 support contract, there's the time for the outage to happen, be diagnosed, get the technician in, get the parts installed . . ." Ingram noted. "You've got potentially a very long mean time to recovery", and that translates to lost revenue.
Going to the cloud won't eliminate all your worries. "Hardware is still your problem. A server down is still a server down." That's doubly the case when you integrate private cloud with public instances — a useful strategy but one you need efficient systems to manage.
Failure comes in many shapes. Individual 'instance' failures (where a single virtual machine collapses) can be relatively easy to recover from. Underlying server hardware failures by your provider are harder to tackle. "You cannot shut down or reboot your instance in that case. There is no guarantee of when you're going to get that server back," Ingram said.
Don't create an overly complex environment. While you can access almost any infrastructure service via a cloud provider, that doesn't mean you should. "Keep it simple: complexity doesn't scale," Ingram said.
Redundancy is key. Obvious but worth reinforcing. "Have two of everything and make sure you put it in different data centres," Ingram suggested.
Monitor everything. "It's better to have too much monitoring than not enough," Ingram said. Track as many diagnostics as possible, and understand what they mean to your operations.
Automate and write scripts whenever you can. "The key to working with the cloud is making sure that you've got automation down pat," Ingram said. The more you can script, the more you can automate and devolve.
Lifehacker's World Of Servers sees me travelling to conferences around Australia and around the globe in search of fresh insights into how server and infrastructure deployment is changing in the cloud era. This week, I'm in Canberra for Linux.conf.au, paying particular attention to the server administration mini-conference and sessions on virtualisation and best practice.