Date: Fri, 15 Feb 2008 09:48:03 -0700 (MST) From: Brett Bump <bbump@rsts.org> To: Mike Tancsa <mike@sentex.net> Cc: Kris Kennaway <kris@freebsd.org>, freebsd-performance@freebsd.org Subject: Re: System perforamance 4.x vs. 5.x and 6.x Message-ID: <20080215085714.K79197@mail.rsts.org> In-Reply-To: <200802151440.m1FEeGVr084431@lava.sentex.ca> References: <20080214114759.R75215@mail.rsts.org> <47B49A16.1080103@FreeBSD.org> <20080214131026.Y75492@mail.rsts.org> <200802151440.m1FEeGVr084431@lava.sentex.ca>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, 15 Feb 2008, Mike Tancsa wrote: > >Dell PowerEdge 1750 1U, 146Gig U320s. The Broadcoms seem to be a change > >from the earlier 1550s with intel pro/100s (I prefer the intel's). > > So this is not the same hardware as before that was running releng_4 ? Yes, it is actually the same physical box. The mail server started life on a PowerEdge 1550 1U with intel nics, but was rapidly running out of storage. We had a backup 1750 for the content management system so put 4.10 on it and then just copied the users (about 8k of them) and their files across. The 4.x line was considered grossly out of date at the time, but was the quickest and easiest way to do an upgrade in under 1 hour without the users calling wondering where the mail server had gone. I still have 18 4.10 machines out in the field running various other tasks, 24/7 365+ a year spamming me daily information as to the health of my networks. But this one has me baffled as the machine seems perfectly sane for 18 hours a day and can then turn into a slug from 10am-4pm (peak user activity). > >I was suspicious that maybe we needed more memory but swap has barely even > >been touched (232k used...with 1400meg inactive). > > Stiill, it might help by allowing more caching... Also I would still > increase dirhash as you are getting close to the limit. Also, if you > have a large master.passwd file (e.g. > 1000), try changing > nsswitch.conf as instructed in > > http://www.freebsd.org/cgi/query-pr.cgi?pr=75855 > > We had to do this on our pop server otherwise just doing an ls in > /var/mail would take several minutes. I know that one quite well. ;-) I did a massive user cleanup/renumber last summer and blew away about 5k users. I'm now sitting with about 3k and we run 10 different groups as well as the groups added with clamav, mailman, mysql, www etc. The 4.x versions never caused me to wait on a directory listing, however with so many files you can't wildcard search anything. I presumed the change to UFS2 had something to do with this. I'll increase dirhash and have already switched nsswitch.conf to files. > I seem to recall people having issues with the media selection using > bge based nics. e.g. > http://www.freebsd.org/cgi/query-pr.cgi?pr=112570 > > I would try using autoneg instead. There are other options that might > not be getting set right (e.g. FC). autoneg might take care of it > for you, but as I said before there have been a number of bug fixes > to the driver since 6.2. Similarly, the ports you are using have > known security issues from 6.2 so you are better off to start from > 6.3 and its port as you will have less patching to do. > > ---Mike > I started with autoneg and changed it because I thought maybe we were having network congestion problems, but I've very sure that is now not the case (so yes, this one can easily be changed back). Possilby the weekend project will be setting up the backup with 6.3 to switch everything over to that. Brett
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20080215085714.K79197>