Hi all, I downloaded a pre-compiled Fedora 11 domU image from http://stacklet.com/downloads/images/list/Fedora When I ran the domU (2.6.30) with xen 3.3.1, 2.6.18 dom0, there was a problem. I can not change the memory configuration of the domU on the fly. I used pygrub to boot the PV guest. In the 2.6.30 guest, I can not find the /proc/xen/balloon driver but in the config file of the kernel, XEN_BALLOON is enabled. Any ideas? Thanks. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On 11/29/09 21:36, Jia Rao wrote:> Hi all, > > I downloaded a pre-compiled Fedora 11 domU image from > http://stacklet.com/downloads/images/list/Fedora > > When I ran the domU (2.6.30) with xen 3.3.1, 2.6.18 dom0, there was a > problem. > > I can not change the memory configuration of the domU on the fly. I > used pygrub to boot the PV guest. > In the 2.6.30 guest, I can not find the /proc/xen/balloon driver but > in the config file of the kernel, XEN_BALLOON is enabled.How were you trying to change the memory? At present, it only supports shrinking memory, not expanding it beyond its original size. J _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
OK, the shrinking is working now. What I do now is to set a large initial memory size. Another problem is with the newest pv_ops kernel. I downloaded the xen-unstable, and compiled it under fedora 10. Dom0 booted fine, but when I cp the dom0 to a domU (what I did in the stable xen release, convert physical to virtual), the domU can boot, but immediately the vcpu(i set to 2) shrinked to 1 vcpu and the cpu was 100% busy there. ping and xm console returned nothing. The same thing happened when I install a precompiled pv_ops kernel to my old domUs. I tried xen 3.3.1 (2.6.18.8 dom0) and xen-unstable(2.6.31.6 dom0), same problem. So I think the problem should be the pv_ops kernel. Do I need some special configuration for pv_ops kernel to use as a domU? Thanks ! On Mon, Nov 30, 2009 at 4:08 PM, Jeremy Fitzhardinge <jeremy@goop.org>wrote:> On 11/29/09 21:36, Jia Rao wrote: > > Hi all, > > > > I downloaded a pre-compiled Fedora 11 domU image from > > http://stacklet.com/downloads/images/list/Fedora > > > > When I ran the domU (2.6.30) with xen 3.3.1, 2.6.18 dom0, there was a > > problem. > > > > I can not change the memory configuration of the domU on the fly. I > > used pygrub to boot the PV guest. > > In the 2.6.30 guest, I can not find the /proc/xen/balloon driver but > > in the config file of the kernel, XEN_BALLOON is enabled. > > > How were you trying to change the memory? At present, it only supports > shrinking memory, not expanding it beyond its original size. > > J >_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Konrad Rzeszutek Wilk
2009-Nov-30 21:50 UTC
Re: [Xen-devel] Problem with 2.6.30 pv-ops domU
> Do I need some special configuration for pv_ops kernel to use as a domU?If you have ''fb'' in your .xm file you are hitting a bug that I am trying to fix. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Can you elaborate how can I verified if it is the bug, what .xm file you are referrring? On Mon, Nov 30, 2009 at 4:50 PM, Konrad Rzeszutek Wilk < konrad.wilk@oracle.com> wrote:> > Do I need some special configuration for pv_ops kernel to use as a domU? > > If you have ''fb'' in your .xm file you are hitting a bug that I am trying to > fix. >_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Konrad Rzeszutek Wilk
2009-Nov-30 23:05 UTC
Re: [Xen-devel] Problem with 2.6.30 pv-ops domU
On Mon, Nov 30, 2009 at 04:55:24PM -0500, Jia Rao wrote:> Can you elaborate how can I verified if it is the bug, what .xm file you are > referrring?This: [root@tst002 xen]# cat /etc/xen/fc11.xm |grep fb vfb=[''type=vnc,vnclisten=0.0.0.0,vncunused=1''] Irregardless of the ''vfb'' arguments. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
No, I did not specify an argument for framebuffer. On Mon, Nov 30, 2009 at 6:05 PM, Konrad Rzeszutek Wilk < konrad.wilk@oracle.com> wrote:> On Mon, Nov 30, 2009 at 04:55:24PM -0500, Jia Rao wrote: > > Can you elaborate how can I verified if it is the bug, what .xm file you > are > > referrring? > > This: > > [root@tst002 xen]# cat /etc/xen/fc11.xm |grep fb > vfb=[''type=vnc,vnclisten=0.0.0.0,vncunused=1''] > > Irregardless of the ''vfb'' arguments. >_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On 11/30/09 13:44, Jia Rao wrote:> OK, the shrinking is working now. What I do now is to set a large > initial memory size. > > Another problem is with the newest pv_ops kernel. > > I downloaded the xen-unstable, and compiled it under fedora 10. > > Dom0 booted fine, but when I cp the dom0 to a domU (what I did in the > stable xen release, convert physical to virtual), the domU can boot, > but immediately the vcpu(i set to 2) shrinked to 1 vcpu and the cpu > was 100% busy there.What do you mean by "shrinked to 1"?> ping and xm console returned nothing. > > The same thing happened when I install a precompiled pv_ops kernel to > my old domUs. > I tried xen 3.3.1 (2.6.18.8 dom0) and xen-unstable(2.6.31.6 dom0), > same problem. > So I think the problem should be the pv_ops kernel.What happens if you run with one vcpu? What messages are printed when it boots? Any? Do you have "earlyprintk=xen" on the kernel command line? Does it respond to sysrq? Can you get any info about the domain with xenctx?> Do I need some special configuration for pv_ops kernel to use as a domU?No. J _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
The problem was solved. It turned out to be the blktap support in pv_ops. I configured all the VM disks as tap:aio which seems to be waiting for tap:aio 2. After changed to file based disk, everything is fine up to now. On Mon, Nov 30, 2009 at 7:25 PM, Jeremy Fitzhardinge <jeremy@goop.org>wrote:> On 11/30/09 13:44, Jia Rao wrote: > > OK, the shrinking is working now. What I do now is to set a large > > initial memory size. > > > > Another problem is with the newest pv_ops kernel. > > > > I downloaded the xen-unstable, and compiled it under fedora 10. > > > > Dom0 booted fine, but when I cp the dom0 to a domU (what I did in the > > stable xen release, convert physical to virtual), the domU can boot, > > but immediately the vcpu(i set to 2) shrinked to 1 vcpu and the cpu > > was 100% busy there. > > What do you mean by "shrinked to 1"? > > > ping and xm console returned nothing. > > > > The same thing happened when I install a precompiled pv_ops kernel to > > my old domUs. > > I tried xen 3.3.1 (2.6.18.8 dom0) and xen-unstable(2.6.31.6 dom0), > > same problem. > > So I think the problem should be the pv_ops kernel. > > What happens if you run with one vcpu? > What messages are printed when it boots? Any? Do you have > "earlyprintk=xen" on the kernel command line? > Does it respond to sysrq? > Can you get any info about the domain with xenctx? > > > Do I need some special configuration for pv_ops kernel to use as a domU? > > No. > > J >_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel