From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:47230) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1WIJoL-0004sw-B3 for qemu-devel@nongnu.org; Tue, 25 Feb 2014 10:20:52 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1WIJoD-0005Av-35 for qemu-devel@nongnu.org; Tue, 25 Feb 2014 10:20:44 -0500 Received: from mail3-relais-sop.national.inria.fr ([192.134.164.104]:29385) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1WIJoC-00058U-Ei for qemu-devel@nongnu.org; Tue, 25 Feb 2014 10:20:36 -0500 Message-ID: <530CB49D.3020202@inria.fr> Date: Tue, 25 Feb 2014 16:19:57 +0100 From: Vincent KHERBACHE MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Subject: [Qemu-devel] dirty pages refresh List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: qemu-devel@nongnu.org Hi all, By playing with dirty pages logging, I observed that the amount of dirty pages increase very quickly just after a bitmap cleaning, and then more slowly. I guess that it is probably due to the workload which mostly realloc the same block of memory. So, in this case, it means that the dirty pages rate mainly depends on the frequency of dirty bitmap cleaning+syncing, is that right ? If yes, I assume that during a live migration the dirty pages rate directly depends on the available bandwidth (dirty page sent => bitmap bit cleaning), or am I wrong once again ? Hope your helps, Regards. -- Vincent KHERBACHE