Date: Wed, 26 Apr 2000 11:02:15 -0400 (EDT) From: Marc Tardif <intmktg@CAM.ORG> To: Shawn Barnhart <swb@grasslake.net> Cc: freebsd-questions@FreeBSD.ORG Subject: Re: clustering and netbooting Message-ID: <Pine.LNX.4.10.10004261054240.11556-100000@Gloria.CAM.ORG> In-Reply-To: <028f01bfaf90$40941060$b8209fc0@marlowe>
next in thread | previous in thread | raw e-mail | index | archive | help
> | Problem is: how can I avoid having to build/install world and, to a lesser > | extent, the kernel for each machine (same hardware)? > | > | Solution 1: netbooting > [snip] > | Solution 2: mounting directories from fstab > [snip] > > I've done variations of both, netbooting and nfs mounting key directories, > and while its a great solution for end-user type workstations, non-critical > servers or identical server farms, I've found that by focusing my points of > failure to one machine I increase my troubles, not decrease them when > dealing with critical services. > > Even if you have a rock-solid NFS setup, what happens when a make > installworld screws up? Do you want to lose all the machines, or just one > of them? Right, that's why I currently only buildworld on one machine and nfs mount /usr/src and /usr/obj on the other machines. Although I only have to installworld and rebuild the kernel once for each machine, this can become tedious for a farm of 10+ machines. Is there perhaps a way to automate the task of keeping the system up to date? Yahoo! probably has quite a large cluster of machines, do they keep up to date one machine at the time? Or how about this: could two machines act as boot servers for all the other machines? That way, if one fails, the other takes over provinding somewhat of a fail-safe environment. Not sure about that though... To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.LNX.4.10.10004261054240.11556-100000>