tag:blogger.com,1999:blog-4957646248778099442.post8428374648669868727..comments2023-06-30T07:51:45.252-07:00Comments on dev2ops: delivering application change: Dev sees the world one way while Ops sees it a different wayAlex Honorhttp://www.blogger.com/profile/13700437911627383122noreply@blogger.comBlogger2125tag:blogger.com,1999:blog-4957646248778099442.post-25111563685180802652009-04-04T02:10:00.000-07:002009-04-04T02:10:00.000-07:00Dev and Ops certainly do see the world differently...Dev and Ops certainly do see the world differently. Maybe it's too much to ask for the two groups to work together to support software releases, when dev and ops have essentially opposite incentives: dev wants to release as many new features as they can, while ops strives to maintain stability and uptime.<BR/><BR/>Maybe what's needed is a mediator between the groups--a team that understands the software and the infrastructure. Google calls it Site Reliability Engineering, but I don't think the concept is limited to web shops.<BR/><BR/>If you're interested, I've outlined the case for SRE here:<BR/><BR/><A HREF="http://daemons.net/~clay/2009/04/02/engineering-and-operations-bridging-the-divide/" REL="nofollow">http://daemons.net/~clay/2009/04/02/engineering-and-operations-bridging-the-divide/</A>Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-4957646248778099442.post-2985346082425795582007-08-31T11:55:00.000-07:002007-08-31T11:55:00.000-07:00Application deployment is a good case that shows h...Application deployment is a good case that shows how the differing dev and ops view points guide how they apply their associated toolset to a problem.<BR/>A developer might envision using an SCM to store all the release artifacts, building an automated tool that checks them out (perhaps using a tagging convention tied to release plan) to the target machines, and then using a build tool to finish the installation. They use SCM and build tools to conduct their daily work and understand their strengths so it's no wonder they see them useful outside of their typical context.<BR/>An ops person, on the other hand, may be accustomed to copying files from a central file server to the target machines using tools like rsync for distribution. They may script a utility that relies on a file directory structure of an NFS server to maintain release sets and use sed and awk commands to customize files after distribution. Those are familiar tools and approach to an ops person. <BR/>Just examples of how toolsets can reflect and enforce a view point.Alex Honorhttps://www.blogger.com/profile/13700437911627383122noreply@blogger.com