From mboxrd@z Thu Jan 1 00:00:00 1970 From: Julian Anastasov Subject: Re: [PATCH] Sloppy TCP, SH rebalancing, SHP scheduling Date: Thu, 20 Jun 2013 23:09:16 +0300 (EEST) Message-ID: References: <20130524120935.GL264@eldamar.org.uk> <20130524151408.GM264@eldamar.org.uk> <519F92EB.4080509@gmail.com> <51A4B3ED.4070809@gmail.com> <51C335B6.2000202@gmail.com> Mime-Version: 1.0 Return-path: In-Reply-To: <51C335B6.2000202@gmail.com> Sender: lvs-devel-owner@vger.kernel.org List-ID: Content-Type: TEXT/PLAIN; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Aleksey Chudov Cc: lvs-devel@vger.kernel.org Hello, On Thu, 20 Jun 2013, Aleksey Chudov wrote: > Hi, > > I'm trying to calculate how much memory will be needed to synchronize the > persistent templates across the entire cluster. How much memory consumes one > persistent template? Just like every connection, may be you can find such message in dmesg/logs when IPVS starts: IP_VS_DBG(0, "Each connection entry needs %Zd bytes at least\n", sizeof(struct ip_vs_conn)); You can also check 'slabtop' (cat /proc/slabinfo) on load, look for ip_vs_conn. On my x86-32 platform it shows 240 bytes. On x86-64 it should be more. Regards -- Julian Anastasov