I''m trying to start a number of lvm-backend domUs on a Dell Poweredge 2850 (i''ve got nousb on the grub command line, and no usb in the kernel). Some domUs are created and started successfully. However, quite a few of my domUs don''t start properly. I get the correct Started Domain messages... but then the domain doesn''t "go anywhere", meaning it uses .3 cpu time, then never any more. I create the domUs by dd if=/dev/xen/base_image of=/dev/xen/ new_image, after creating a new_image using lvcreate of the same size as the base image. I''ve verified that these images actually exist. Looking through the log from xend.log below, i see an error about / dev/hda1 (the root device given to the domU on the command line, which you can also see in the logs) not existing. Why would that be? I''m running the xen-3.0 testing... The physical machine has 2 Xeon dual-core processors with hyperthreading (which is enabled). Should I disable hyperthreading in the kernel? Would that make a difference? I saw a post earlier about that... Error log attached, with ip addresses replaced. Thanks for the help. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
On Jan 24, 2006, at 9:38 AM, Luke wrote:> I''m trying to start a number of lvm-backend domUs on a Dell > Poweredge 2850 (i''ve got nousb on the grub command line, and no usb > in the kernel). > > Some domUs are created and started successfully. > > However, quite a few of my domUs don''t start properly. I get the > correct Started Domain messages... but then the domain doesn''t "go > anywhere", meaning it uses .3 cpu time, then never any more. > > I create the domUs by dd if=/dev/xen/base_image of=/dev/xen/ > new_image, after creating a new_image using lvcreate of the same > size as the base image. I''ve verified that these images actually > exist. > > Looking through the log from xend.log below, i see an error about / > dev/hda1 (the root device given to the domU on the command line, > which you can also see in the logs) not existing. Why would that be? > > I''m running the xen-3.0 testing... > > The physical machine has 2 Xeon dual-core processors with > hyperthreading (which is enabled). Should I disable hyperthreading > in the kernel? Would that make a difference? I saw a post earlier > about that... > > Error log attached, with ip addresses replaced. > <xen.log> > > Thanks for the help.Turns out, the problem seems even more serious. All long running jobs get unexpectedly terminated. So I couldn''t compile a new Xen kernel, since it kept getting interrupted - even screen was getting killed. Anyone else have similar problems? Any insight as to what the problem could be? I wasn''t having any issues on a 1850 Poweredge w/ similar processors and hyperthreading enabled, but on this one, everything dies (all domUs turn to Zombies after a bit...) _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
[snip]> Turns out, the problem seems even more serious. All long running > jobs get unexpectedly terminated. So I couldn''t compile a > new Xen kernel, since it kept getting interrupted - even > screen was getting killed. > > Anyone else have similar problems? > > Any insight as to what the problem could be? I wasn''t having > any issues on a 1850 Poweredge w/ similar processors and > hyperthreading enabled, but on this one, everything dies (all > domUs turn to Zombies after a bit...) >What happens if you try to build Xen under a native Linux? Have you tried running something like memtest86 to verify that the machine is actually working OK? -- Mats _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
On Jan 25, 2006, at 9:26 AM, Petersson, Mats wrote:> What happens if you try to build Xen under a native Linux? >Building / every other long process works just fine under native Linux. In fact, it works under Xen eventually, after starting it up numerous times (it gets a bit further every time).> Have you tried running something like memtest86 to verify that the > machine is actually working OK?Not yet, but under debian''s stock kernel, I wasn''t seeing any odd behaviors. I''ll run some tests and see. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users