Date: Wed, 15 Oct 97 13:10:11 -0700 From: "Studded" <Studded@dal.net> To: "David Greenman" <dg@root.com>, "Tom" <tom@uniserve.com> Cc: "Matthew M Groener" <matt@xlrn.ucsb.edu>, "stable@FreeBSD.ORG" <stable@FreeBSD.ORG> Subject: Re: out of mbuf crash Message-ID: <199710152010.NAA22836@mail.san.rr.com>
next in thread | raw e-mail | index | archive | help
On Wed, 15 Oct 1997 10:35:11 -0700 (PDT), Tom wrote: > >On Wed, 15 Oct 1997, Studded wrote: > >> nmbclusters kernel option. I've raised it to 15000 on a 2.2.1 system. > > nmbclusters are statically allocated. Pre-allocating 15000 will consume >a LOT of RAM. As Dave pointed out, this isn't exactly true, however if you're running a very high capacity system, this is one of the things you have to take into account. With 4,000 users we regularly run with 8-9,000 mbufs in use, so I need the overhead. We also have 128M of ram, so that's not really an obstacle. >> 1383 mbufs in use: >> 649 mbufs allocated to data >> 725 mbufs allocated to packet headers >> 8 mbufs allocated to protocol control blocks >> 1 mbufs allocated to socket names and addresses >> 62/3574 mbuf clusters in use >> >> The number after the / in the 6th line will tell you >> (basically) the highest number of mbuf's ever requested (although I've >> seen higher numbers in use than the max on the 6th). You want the >> number of mbuf's available to be 50% greater than the max requested. > > The number after the "/", is the number of mbufs currently allocated, >there may never have that many mbufs in use. When the number of mbufs in >use reaches the number of mbufs allocated, another whole chunk of mbufs >are allocated (assuming you have not reached the limit set by >NMBCLUSTERS). Thank you for clarifying that. :) Doug *** Proud operator, designer and maintainer of the world's largest *** Internet Relay Chat server. 4,168 clients and still growing. :-) *** Try spider.dal.net on ports 6662-4 (Powered by FreeBSD)
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199710152010.NAA22836>