From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ian Campbell Subject: Re: null domains | xen4.4 Date: Mon, 1 Dec 2014 11:00:05 +0000 Message-ID: <1417431605.29138.5.camel@citrix.com> References: <1417427958.27655.1.camel@citrix.com> <2a9dd806b9991d506ae274ff67fe5821@alukardd.org> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <2a9dd806b9991d506ae274ff67fe5821@alukardd.org> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xen.org Errors-To: xen-devel-bounces@lists.xen.org To: Alexey Cc: Xen-devel@lists.xen.org List-Id: xen-devel@lists.xenproject.org On Mon, 2014-12-01 at 13:54 +0300, Alexey wrote: > xl list two domains for example: > (null) 67 0 8 --pssd > 135229.6 > (null) 69 0 4 --ps-d > 4172.2 > > vif67.0 still exist in my dom0 and i can't "ip link delete" it. > Attached block device marked as busy also. What does "xenstore-ls -fp" say? Are the backend xenstore nodes still present? > We try to use the last 3.10 kernel, now new nodes use 3.10.61 kernel, > but this node on which I can experiment was booted with 3.10.55 kernel. I meant something newer than 3.10, not just the latest 3.10. > I attach xl-dmesg after 'q' trigger. Thanks, seems like each domain has three leaked pages, e.g.: (XEN) Memory pages belonging to domain 4: (XEN) DomPage 0000000000c2b7e6: caf=00000001, taf=7400000000000001 (XEN) DomPage 0000000000c2aa03: caf=00000001, taf=7400000000000001 (XEN) DomPage 0000000000cb8397: caf=00000001, taf=7400000000000001 If the vifX.Y device is still in existence then there's a pretty good chance it's holding a page, likewise any disk backends etc. Are there any messages in the dom0 dmesg around the time of the guest supposedly shutting down? How about logs under /var/log/xen. Ian.