From owner-freebsd-questions Mon Dec 7 12:16:48 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id MAA10751 for freebsd-questions-outgoing; Mon, 7 Dec 1998 12:16:48 -0800 (PST) (envelope-from owner-freebsd-questions@FreeBSD.ORG) Received: from Central.KeyWest.MPGN.COM (Central.TanSoft.COM [12.16.39.1]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id MAA10746 for ; Mon, 7 Dec 1998 12:16:47 -0800 (PST) (envelope-from rwm@tansoft.com) Received: from devious (Devious.TanSoft.COM [12.16.39.10]) by Central.KeyWest.MPGN.COM (8.9.0/8.9.0) with SMTP id PAA21898; Mon, 7 Dec 1998 15:16:20 -0500 Message-Id: <4.1.19981207150722.00933230@central.TanSoft.COM> X-Sender: rwm@central.TanSoft.COM X-Mailer: QUALCOMM Windows Eudora Pro Version 4.1 Date: Mon, 07 Dec 1998 15:15:13 -0500 To: aron From: Rob Miracle Subject: Re: NMBCLUSTERS: Max Sane Values?? Cc: FreeBSD-questions@FreeBSD.ORG In-Reply-To: Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG At 10:52 AM 12/6/98 -0800, you wrote: >The handbook says max for NMBCLUSTERS is 4096 however I am pretty sure >that it is out of date... We are running 2.2.7 on our NFS server and we reboot almost daily with an out of mbufs error. We have maxusers defined as 128 and the param.c file is defining NMBCLUSTERS as 512 + maxusers * 16 and nmbufs = NMBCLUSTERS * 4, which results in 2560 clusters and over 10,240 nmubs. >I currently have NMBCLUSTERS set to 10240 with no ill effects yet i still >get panics every two days or so... (the box mostly runs tons of apache >procs) >and why does this situation cause a kernel panic anyway?? I seem to >remember a similar situation with NetBSD a while back which cured the >kernel panics when mbufs were exhausted and thought freebsd had also >corrected the problem even earlier.... > >any info or pointers appreciated. netstat -m reports far fewer mbufs in use that we think we have and it shows a large amount of kbyte usage: 346 mbufs in use: 265 mbufs allocated to data 73 mbufs allocated to packet headers 6 mbufs allocated to protocol control blocks 2 mbufs allocated to socket names and addresses 165/202 mbuf clusters in use 447 Kbytes allocated to network (83% in use) 0 requests for memory denied 0 requests for memory delayed 0 calls to protocol drain routines We have been having period panics in 2.2.7 and earlier with an out of mbufs error on our main NFS server which is also running the database and another data server. This is starting to drive me nuts and management is getting a bit upset about the downtime and they are throwing other OS's around (hissssss). I need some help on this. Rob To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message