From owner-freebsd-hackers Mon Feb 14 9:43:25 2000 Delivered-To: freebsd-hackers@freebsd.org Received: from brunel.uk1.vbc.net (brunel.uk1.vbc.net [194.207.2.8]) by builder.freebsd.org (Postfix) with ESMTP id 1D4924C75; Mon, 14 Feb 2000 09:43:20 -0800 (PST) Received: from localhost (lloyd@localhost) by brunel.uk1.vbc.net (8.9.3/8.9.3) with ESMTP id RAA40840; Mon, 14 Feb 2000 17:43:36 GMT Date: Mon, 14 Feb 2000 17:43:36 +0000 (GMT) From: Lloyd Rennie To: freebsd-hackers@freebsd.org, freebsd-net@freebsd.org Subject: GateD returns 'No buffer space available' upon a routing table build... Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG Hey all; We have several routers storing a full global routing table. Recently, this table has grown rather large (currently at ~75k routes). All are behaving in the same way; When GateD starts, I can watch it growing on a 'top', as it builds a routing table. Suddenly, it stops growing, cpu usage drops to nill, and many kernel messages are spewed-out; Feb 14 13:47:34 gate3 /kernel: arp_rtrequest: malloc failed Feb 14 13:47:34 gate3 /kernel: arplookup 194.207.0.32 failed: could not allocate llinfo Feb 14 13:47:34 gate3 /kernel: arpresolve: can't allocate llinfo for 194.207.0.32rt Feb 14 13:47:34 gate3 gated[88]: KRT SENT type ADD(1)flags UP GW(3) error 55: Nobuffer space available Feb 14 13:47:34 gate3 gated[88]: KRT SENT dest 204.0.41 gateway 194.207.0.46 netmask 255.255.255 Feb 14 13:47:37 gate3 /kernel: arplookup 194.207.0.32 failed: could not allocate llinfo etc... The first symptom seems always to be the malloc failure. This happens within the first minute of uptime on this particular machine, a p2-400/128Mb. All machines run FreeBSD-3.2. The others take longer to develop the fault, but have slower processors. My assumption is that the global routing table has grown to such a size that there is no longer enough kernel memory available to store it. I checked through the LINT file, but couldn't find any useful options. Started looking tentatively through the kernel source, but (bear in mind I'd rather not lower maxusers anymore) came to a dead end with a 0-byte file called vm_opt.h :( So I'd like to ask... * Which particular type of 'buffer space' is this? * What can I do to 'tweak' it (ie. assign more) * Is the issue resolved in a later release of FreeBSD (pref in the STABLE tree, which reminds me - when's the codefreeze on v4 end?) Any help with this issue would be appreciated... -- Lloyd Rennie VBCnet GB Ltd lloyd@vbc.net tel +44 (0) 117 929 1316 http://www.vbc.net fax +44 (0) 117 927 2015 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message