Date: Fri, 22 Dec 2006 00:15:13 -0500 From: "Michael R. Wayne" <freebsd@wayne47.com> To: stable@freebsd.org Subject: Re: Possibility for FreeBSD 4.11 Extended Support Message-ID: <20061222051513.GK63341@manor.msen.com> In-Reply-To: <200612220259.kBM2xYxc019408@khavrinen.csail.mit.edu> References: <000801c723bb$efc2b540$260ba8c0@wii.wintecind.com> <458A2B14.5070009@freebsd.org> <458A97BF.1090503@ant.uni-bremen.de> <20061221095811.886d9850.wmoran@collaborativefusion.com> <200612220259.kBM2xYxc019408@khavrinen.csail.mit.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, Dec 21, 2006 at 09:59:34PM -0500, Garrett Wollman wrote: > spork@bway.net writes: > > >-5.x was never really for production use, in the same way 3.x never > >was. > > Why do people continue to say this? Many sites have used, are still > using, and plan to continue to use, 5.x in production. I'm going to copy a bit of mail that I sent to someone privately. FreeBSD 4.11 can survive a simple burn-in test. FreeBSD 5.X and 6.1 can not. Here's what I wrote earlier. Take a server. Configure for SMP, add quotas within jails and basic IPFW protection with a few hundred dummynet pipes for b/w throttling (less than 10,000 total IPFW lines). Load the machine a bit so that it constantly maintains a 3 digit load and run sufficient active processes to keep it in moderate swap state. The result of that minimal-effort test yields machines which can not maintain 30 days of uptime (most fail in under a week). And don't even THINK about snapshots in 6.1 or earlier. >THAT< is why people who run servers, with jails, quotas, ipfw and moderate load keep complaining about 5.X and 6.1 and begging for 4.11 support to be extended. Just because someone has a few FreeBSD boxes running light loads and not using the features that we NEED does not mean that any the port 4.11 releases to date are stable. /\/\ \/\/
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20061222051513.GK63341>