Date: Wed, 14 Aug 2002 10:32:04 -0400 From: Matthias Trevarthan <trevarthan@wingnet.net> To: freebsd-stable@freebsd.org Subject: Re: setting up a CVSup repository Message-ID: <200208141032.04447.trevarthan@wingnet.net> In-Reply-To: <20020814140526.GA29078@tp.databus.com> References: <200208140919.35737.trevarthan@wingnet.net> <20020814140526.GA29078@tp.databus.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wednesday 14 August 2002 10:05 am, Barney Wolff wrote: > This can be done, as another poster has indicated. But it may be too > much effort for what you want to accomplish. There are multiple ways > to administer a collection of FreeBSD systems without having each one > do its own cvsup: > Nope. This is exactly what I want. > 1. As you asked, set up your own cvsup mirror. It seems to me that > this is the way to go only if the systems that will be using it are > not under your direct supervision. > > 2. Do cvsup of the cvs archive on one machine, then have others do > their own remote CVS checkouts from the archive on that. This is simpler > in some ways on the server, and really no harder on the clients. It > allows you to build current and stable and cpu flavors, as you wish. > I'm a little confused about the differences between one and two. Could you elaborate? > 3. cvsup on one machine, build on that, and have all the others > NFS mount /usr/src, /usr/obj and /usr/ports. This has the feature > that you control which version is in use and saves a lot of time on > all the client machines. It is clearly the way to go if all the > machines are under your supervision and you're willing to build > stuff that will run on all your cputypes - the optimizations available > for each type are really minor within the x86 family so the loss > of the last inch of performance is worth the generality, imho. I > build separate kernels for each x86 flavor but a common world. > I'm not real fond of NFS. We have a somewhat distributed network, with multiple server rooms connected by T1s. They're all under my administration, but I think running CVSup on each machine is fairly ideal. UNIX machines multitask pretty well, and it only takes about two hours to make buildworld on my beefy servers. I may do this within server rooms though.. Have one master server download and build the world in each room, then distribute via NFS inside each room... Don't know. I'll have to think about it. > I actually do #2 but only do the checkout on the local machine and > build there. Again, how does this differ from #1? To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200208141032.04447.trevarthan>