I've gone through the whole project Mercury setup on one of my personal non-aws servers (i386). I've come to some conclusions and i would like some feedbacks :
-
Seems The project Mercury is attempting to do some "automation" for a whole server stack (packages, configurations, applications...) for your drupal setup
-
If you can actually setup apache, mysql, solr, varnish... do you think you still need to check project Mercury?
-
Seems the project plays with packages, hence possibly system packages (install/remove)... and that could cause some serious problems if you're managing your own server, already have your server setup and configurations and just want a fast stack configuration for your drupal setup.
-
are there ways to just extract "mercury special configurations" and apply them on top of previous varnish + apache + mysql + memcache + solr setup ?
I would love some good feedback on this.
Thanks in advance
Comments
Sort of
The "optimal stack" is a moving target. The mercury AMIs were a way for us to start getting these concepts out there and prove the validity of BCFG2 as a management platform. We assume you'll be devoting the server to Drupal.
If you can actually setup apache, mysql, solr, varnish... do you think you still need to check project Mercury?
One question is "do you want to be a system configurator or a drupal developer". It's a lot of work to be really highly skilled at both, and there's a lot more project-by-project nuance to Drupal sites than there is to a great Drupal hosting stack. Our vision is that people can run high performance Drupal servers without having to keep up with all the relevant details.
Another question is "do you want to spend time typing apt-get and messing with conf files". If you do this a lot it gets repetitive. That's one of the reasons we bothered to create this project. ;)
are there ways to just extract "mercury special configurations" and apply them on top of previous varnish + apache + mysql + memcache + solr setup ?
The whole stack is open-source, so you are free to extract all the configurations from the repository and use them as a guideline for setting up the server.
https://pantheon.io | http://www.chapterthree.com | https://www.outlandishjosh.com
Thanks for the explanation
I was just a bit skeptical seeing things "auto magically" being setup and removed with bcfg2. And it's a bit scary on servers that you've already setup without mercury. It's also another story if you already have your server architecture structure setup and working before mercury in the hands of a dedicated sys admin.
I'll go deep on bcfg2 and hudson for more information.
Thanks