From mboxrd@z Thu Jan 1 00:00:00 1970 From: Alex Bligh Subject: Re: Fatal crash on xen4.2 HVM + qemu-xen dm + NFS Date: Wed, 16 Jan 2013 18:14:59 +0000 Message-ID: <031B4703B204494BD5204941@Ximines.local> References: <5B4525F296F6ABEB38B0E614@nimrod.local> <50CEFDA602000078000B0B11@nat28.tlf.novell.com> <3B1D0701EAEA6532CEA91EA0@Ximines.local> Reply-To: Alex Bligh Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; Format="flowed" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: Content-Disposition: inline List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xen.org Errors-To: xen-devel-bounces@lists.xen.org To: Stefano Stabellini Cc: Alex Bligh , Konrad Rzeszutek Wilk , Ian Campbell , Jan Beulich , Xen Devel List-Id: xen-devel@lists.xenproject.org Stefano, --On 16 January 2013 17:39:59 +0000 Stefano Stabellini wrote: > Let me elaborate on this: the guest is a PV on HVM guest, so at the very > least the bootloader is going to use the emulated IDE interface to grab > Xen and the kernel. > After that the kernel should use the PV disk interface straight away (I > actually downloaded the image and tried it myself: it is using the PV > disk interface indeed). Indeed. Thanks for taking the time to download the image and try it. Did you see the bug? (using the snapshot and the backing file) > The bug should occur after the guest has switch over to the PV disk > interface (state = 4 on xenstore for the vbd device). It can't be the > IDE emulator triggering the issue. I would agree. It appears to be triggered at exactly the point the guest image rewrites the partition table then tries to resize the filing system. If you fail to change the grub command line, then it hangs about waiting to retrieve cloud metadata from 169.254.169.254 and nothing crashes until this times out. -- Alex Bligh