Date: Thu, 20 Jul 2000 15:24:14 -0700 From: Kent Stewart <kstewart@urx.com> To: stable@FreeBSD.ORG Subject: Re: CVSup server loads (was: Stable broken) Message-ID: <39777C0E.C44DC92F@urx.com> References: <Pine.BSF.4.21.0007200001080.84615-100000@freefall.freebsd.org> <3976A60C.BE9DE986@urx.com> <20000720095625.A91025@hdroam.ssd.loral.com> <3977334E.AE1F303F@urx.com> <200007202101.OAA17106@vashon.polstra.com>
next in thread | previous in thread | raw e-mail | index | archive | help
John Polstra wrote: > > In article <3977334E.AE1F303F@urx.com>, Kent Stewart > <kstewart@urx.com> wrote: > > > That is where I finally got my cvsup update from. I thought cvsup8 > > was much slower than cvsup7 and then I found later that it wasn't > > very fast on either site. > > If you think this is bad then you weren't around in the days when we > used sup. :-) My introduction was cvsup_16, I think. > > Tagging makes everybody's updates take longer, and as a result the > server's client counts go up. CVSup7 has a limit of 20, and I can > tell you that machine is breathing pretty hard. But it seems to be > performing quite well given the load. This is the first time I've > ever seen it max out. I tried for about 15 minutes and then tried cvsup8. It wasn't full but it wasn't fast. You weren't the only site that was maxed out. There must have been a quiet herd waiting for 4.1-RC to work. Kris, and et. al. around 2300-0100 PDT would commit a fix and boom - the cvsup servers were all busy. It wouldn't work and the next patch would go in and each time the servers were full again. I couldn't get in and so all I could do was chuckle and think of fingers on the triggers and who would be first :). > > Other than the increased load because of the number of clients, the > bottleneck usually isn't in the server when the tree is tagged. The > bottleneck is almost always your hard drive in that case, since it has > to edit every file in your repository, write the edited version to a > temp file, check the MD5 signature, and then move the temp file to the > right place again. Two nights ago, when nametoaddr.c was giving us a compile problem, I had a cvsup7 run in 90 seconds. Last night it was taking 20 minutes. I thought my Bronze+ DSL had dropped a ways. Kent -- Kent Stewart Richland, WA mailto:kbstew99@hotmail.com http://kstewart.urx.com/kstewart/index.html FreeBSD News http://daily.daemonnews.org/ Bomber dropping fire retardant in front of Hanford Wild fire. http://kstewart.urx.com/kstewart/bomber.jpg 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?39777C0E.C44DC92F>