Re: [arch] Clarification on How the Rolling Release System Works

[Doug Jolley]

Thanks again for all the thoughts.  It occurs to me that one never really *HAS* to update.  There's no-one holding a gun to your head.  So, I'm thinking that maybe *A* way to proceed is to maintain a test server that has all of the packages installed that we would normally install on a production server.  As packages are updated I can install them on the test server and once I feel good about that I can then install them on the production servers.  The goal is to try to keep all the servers on Current as far as is practical but do so in an orderly manner.  That would give me the option of saying at any point (for whatever reason) that I just really don't want to update this particular package right now.  So, in a sense, it's a bit like maintaining my own repo as has been suggested.  I have a feeling that instances of not wanting to update a particular package for an extended period would be pretty darn rare.

Also it occurs to me that what we're talking about here is how to get from A to B.  In one case the transition is going to be made by pretty much mainting A for an extended period of time with only bug fixes and security patches applied in the interim and then upgrading to B in one fell swoop.  The other approach is by making little incremental changes along the way.  In both cases, what happens is that we move from A to B.  The thing is that while the first approach tends to hold things fairly stable over a longer period of time, the reality is that ultimately one is going to have to move to B in one fell swoop and there may be problems with that.

In any event, I think the thing for me to do is to take a couple of my least mission critical servers and give it a shot.  I really want to use Arch because I like it *SO* much better than anything else.

Thanks for the input.

          ... doug

_______________________________________________
arch mailing list
arch@archlinux.org
http://www.archlinux.org/mailman/listinfo/arch