Currently, to make infrastructure HA and scalable using AWS I am using following components.
Ancestor -- Which will be used by base-image to launch new instances during high traffic.
autoscale -- I have configured auto-scale policies which launch new instances using above autoscaling.
base-image -- Whenever new code-deploy gets launched by instance, autoscale deploy the latest successful revision to that instance.
Best thing is about code-deploy is that if current deployment fails then it will rollback to previous successful revision.
code-deploy -- I have integrated jenkins with code-deploy so that if there is commit on jenkins branch then after running test cases it triggers master .
Note : code-deploy has a base-image kept under cron job. Whenever new instance launched by chef-client it gets registered in autoscale and if I push any chef-server recipes it is automatically deployed to respective nodes through new/updated.
I am thinking it is pretty much good architecture BUT then I come across chef-client .
Will I be able to achieve pretty much same functionality with opswork OR should I go with my current setup ?
Any suggestions would be appreciated.
Thanks
答案 0 :(得分：2)
opsworks os图像中有一个客户端（包括厨师），但厨师以独奏模式运行（即没有中央服务器）。您几乎可以复制Opsworks所拥有的内容，但是您将放弃一些灵活性和控制权。
我要做的是进一步使用您的设置，并通过使用CloudFormation在每次部署流量并将流量从旧堆栈转移到新堆栈时启动新堆栈。这对于保持测试，预测和产品的多个环境也很棒。