From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([208.118.235.92]:58543) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1SXCNq-0008Sr-QM for qemu-devel@nongnu.org; Wed, 23 May 2012 10:17:56 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1SXCNk-0005Bo-Gd for qemu-devel@nongnu.org; Wed, 23 May 2012 10:17:50 -0400 Received: from mail-ob0-f173.google.com ([209.85.214.173]:63385) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1SXCNk-0005BG-9K for qemu-devel@nongnu.org; Wed, 23 May 2012 10:17:44 -0400 Received: by obbwd20 with SMTP id wd20so13868106obb.4 for ; Wed, 23 May 2012 07:17:41 -0700 (PDT) Sender: fluxion Date: Wed, 23 May 2012 09:17:37 -0500 From: Michael Roth Message-ID: <20120523141737.GA5615@illuin> References: MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline Content-Transfer-Encoding: 8bit In-Reply-To: Subject: Re: [Qemu-devel] The image size of instance VM keeps growing List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: =?utf-8?B?Q2hhcmxlcy5Uc2FpLeiUoea4hea1ty3noJTnqbbnmbzlsZXpg6g=?= Cc: =?utf-8?B?Sm9uYWguV3Ut5ZCz5ZCb5YuJLeeglOeptueZvOWxlemDqA==?= , "qemu-devel@nongnu.org" On Mon, May 21, 2012 at 07:47:49AM +0000, Charles.Tsai-蔡清海-研究發展部 wrote: > > > We created a qcow2 based-image for a VM whose size is big when it is created. When a VM is launched, it is a instance VM image which refers to the based-image. Initially, the image size of instance VM is smaller than that of the based-mage. But we found the image size of the instance VM kept growing, when a VM ran in idle state. We doubted this issue is caused by the Windows swap file(virtual memory) that Windows OS pages the memory contents to it during the run-time. We did disable the swap file setting from Windows but the image size of the instance VM still kept growing. > > > > Ideally, the image size of instance VM is supposed to remain in a smaller size than the based-image. Why does it keep growing in size which is even bigger than that of based-image? Our VM, in fact, did nothing during the idle state. > Do you have automatic indexing turned on in the guest? Not sure of the specifics WRT to NTFS, but indexing might be causing metadata updates (particularly last access time) that might be causing a COW on a lot of blocks in a relatively short amount of time.