From owner-freebsd-questions Wed Jan 12 14: 9:14 2000 Delivered-To: freebsd-questions@freebsd.org Received: from pebkac.owp.csus.edu (pebkac.owp.csus.edu [130.86.232.245]) by hub.freebsd.org (Postfix) with ESMTP id 4651B14D62 for ; Wed, 12 Jan 2000 14:08:33 -0800 (PST) (envelope-from joseph.scott@owp.csus.edu) Received: from owp.csus.edu (mothra.ecs.csus.edu [130.86.76.220]) by pebkac.owp.csus.edu (8.9.3/8.9.3) with ESMTP id OAA23688; Wed, 12 Jan 2000 14:08:19 -0800 (PST) (envelope-from joseph.scott@owp.csus.edu) Message-ID: <387CFB4E.827314D9@owp.csus.edu> Date: Wed, 12 Jan 2000 14:08:14 -0800 From: Joseph Scott X-Mailer: Mozilla 4.7 [en] (X11; I; Linux 2.0.36 i386) X-Accept-Language: en MIME-Version: 1.0 To: Sean Heber Cc: freebsd-questions@FreeBSD.ORG Subject: Re: Are huge file systems bad? References: <002d01bf5d43$845331e0$0a04cfd1@mwci.net> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG Sean Heber wrote: > > I'm setting up a server that needs a lot of storage area online via web and > ftp. Basically, what we're doing is giving our >650 registered developers > ftp access so they do not have to find alternative hosting providers. > > Basically, my question is 2-fold.. > > 1) Are huge file systems bad? I have concatenated two large drives > together using vinum. The resulting file system is 50Gig. > > 2) Is there another way to structure this that would keep file systems > small but be easy to manage if number 1 is bad? > > Our structure for the ftp site has the ftp dir for each user under > /sites/ftp/usr. So like this: > > usr/bdev1 > usr/bdev2 > usr/bdev3 > usr/bdev4 > usr/bdev5 > ... > > As far as I can tell, that makes it very hard to break up all the space into > smaller partitions. So that's why I went the vinum route. However, I've > had conflicting reports that large file systems can be really dangerous. So > I'd like to know for sure. > > Yes, this 50GB will be backed up. But I want to avoid any unnecessary > problems ahead of time. From my reading of the different lists I believe the following is the situation as it stands now : 1. Large file systems themselves are not a problem in BSD. Several people have reported having very large file systems on FreeBSD. ( IE : 100G or so ). 2. File systems are large IDE drives seem to be a problem, several people have reported having problems making these large ( ~30G ) disks one file system. The solution thus far has been to break it up into multiple filesystem. I'm not sure exactly where this limit is, possible around ~25G, maybe less. I'm expecting a ~30G IDE disk with in the next few weeks and will see how big I can make it. The IDE drivers in FreeBSD are being replaced so I don't think there will be much work done towards solving this problem until the new drivers are in -STABLE and can be tested with these large IDE drives. Keep in mind this is just my picture of the situation based on reading the different lists. If you are using SCSI disks then #2 may not apply. Depending on what kind of peak load you are looking at then splitting things up ( via vinum or separate disks ) may buy you some performance also. I'm sure there are someways to make splitting up the dir's on to separate file systems managable. Symlinks? Maybe a different naming scheme, like : usr/a/adev1 usr/a/adev2 usr/b/bdev1 usr/b/bdev2 Anyways, just a thought, if it's not what you are looking for feel free to ignore it :-) -- Joseph Scott joseph.scott@owp.csus.edu Office Of Water Programs - CSU Sacramento To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message