From mboxrd@z Thu Jan 1 00:00:00 1970 From: Patrick Schaaf Subject: Re: a question about ip_conntrack_max Date: Thu, 12 Dec 2002 08:21:43 +0100 Sender: netfilter-devel-admin@lists.netfilter.org Message-ID: <20021212072143.GA19846@oknodo.bof.de> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: netfilter-devel@lists.netfilter.org Return-path: To: jump lee Content-Disposition: inline In-Reply-To: Errors-To: netfilter-devel-admin@lists.netfilter.org List-Help: List-Post: List-Subscribe: , List-Unsubscribe: , List-Archive: List-Id: netfilter-devel.vger.kernel.org Hello, (removed Harald Welte from Cc: - I'm sure he's reading the list, anyway) > i am hacking ip_conntrack. What is your goal? > ip_conntrack_max=8*ip_conntrack_htable_size; > > what's the reason of multiplying ip_conntrack_htable_size by 8? It gives you 8 times as much conntrack entries as you have conntrack hash buckets. Isn't 8 a nice number? What number would you prefer, and why? The kernel has to derive ip_conntrack_max SOMEHOW, and that's the way it does it now. There is another equally arbitrary choice of ip_conntrack_htable_size, which is calculated from the amount of memory in your machine. Again, the kernel has to do SOMETHING in the absence of configuration information. best regards Patrick