I just installed and tried to use the Ubuntu Xen packages, and wondered if anyone out there has had similar problems to below. I''m asking here since I''m 90% sure that the problem isn''t the package, but my setup/configuration instead. Firstly, Ubuntu doesn''t build ext3 support into their kernel. Consequently, I''m passing the same initrd to the domU that is passed to dom0 on the command line. I''m passing this via the config file, obviously. Without the initrd, I get a Kernel panic about mounting the root fs. This makes sense, as ext3 support doesn''t exist. With the initrd, I get to the TLS warning about init, then "Continuing...", then nothing. Any ideas what this could be? Could I be using an incorrect initrd? This seems unlikely... I''ve checked that my /etc/fstab is correct... In my config file: disk = [ ''file:/home/myname/vms/myname.raw,hda1,w''] I''ve set up domUs for around 2 years now, which isn''t to say I couldn''t have made a stupid mistake, but I do get the general idea, and I''m not getting helpful error messages here. Any ideas where to start? _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
> I just installed and tried to use the Ubuntu Xen packages, and wondered > if anyone out there has had similar problems to below. I''m asking here > since I''m 90% sure that the problem isn''t the package, but my > setup/configuration instead. > > Firstly, Ubuntu doesn''t build ext3 support into their kernel. > Consequently, I''m passing the same initrd to the domU that is passed to > dom0 on the command line. I''m passing this via the config file, obviously. > > Without the initrd, I get a Kernel panic about mounting the root fs. > This makes sense, as ext3 support doesn''t exist. With the initrd, I get > to the TLS warning about init, then "Continuing...", then nothing.Passing an initrd is usually the most straightforward way to get things working on modern distros.> Any ideas what this could be? > > Could I be using an incorrect initrd? This seems unlikely... > > I''ve checked that my /etc/fstab is correct... > > In my config file: > > disk = [ ''file:/home/myname/vms/myname.raw,hda1,w'']The domain is probably booting but not sending its console output to the Xen virtual serial console. You could try adding "xencons=tty" to the extra= parameter in your domain. Does that fix things for you?> I''ve set up domUs for around 2 years now, which isn''t to say I couldn''t > have made a stupid mistake, but I do get the general idea, and I''m not > getting helpful error messages here.The console has changed behaviour recently. It used to grab the /dev/tty1 device itself so that Xen-unaware distros would Just Work on bootup. That was a slightly rude thing to do in the kernel, so now it uses the special /dev/xvc0 node (or, on mainline Linux 2.6.23 /dev/hvc0) as the console device. Pointing Linux at the correct device using console=/dev/xvc0 to use should solve your problem and allow you to see the output. *however* you''ll notice I''ve not suggested you do this (at least, initially). This is because you''ll still need to modify /etc/inittab and /etc/securetty in order to be able to start a console on the new /dev/xvc0 and log in. Using the "xencons=tty" should make it revert to the older tty-stealing behaviour which should Just Work. You then have the option of moving over to /dev/xvc0 once you''ve switched the config over. You''ll need to move to the more polite /dev/xvc0 route if you want to use the framebuffer console. Hope that helps. Does this fix things for you? Cheers, Mark -- Dave: Just a question. What use is a unicyle with no seat? And no pedals! Mark: To answer a question with a question: What use is a skateboard? Dave: Skateboards have wheels. Mark: My wheel has a wheel! _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Mark Williamson wrote:> > The domain is probably booting but not sending its console output to the Xen > virtual serial console. You could try adding "xencons=tty" to the extra= > parameter in your domain. Does that fix things for you?I believe this is the case. The time on the domain continues to increase. However, with extra = "4 xencons=tty", I now get the last line to be "Setting system clock" after I get "Couldnt get a file descriptor referring to the console", 6 times. So - it seems to be running, but I can''t get a console to fix the networking. So this approach seems not to work for me.> The console has changed behaviour recently. It used to grab the /dev/tty1 > device itself so that Xen-unaware distros would Just Work on bootup. That > was a slightly rude thing to do in the kernel, so now it uses the > special /dev/xvc0 node (or, on mainline Linux 2.6.23 /dev/hvc0) as the > console device. Pointing Linux at the correct device using console=/dev/xvc0 > to use should solve your problem and allow you to see the output. >Is this done in menu.lst/grub command line?> *however* you''ll notice I''ve not suggested you do this (at least, initially). > This is because you''ll still need to modify /etc/inittab and /etc/securetty > in order to be able to start a console on the new /dev/xvc0 and log in. > Using the "xencons=tty" should make it revert to the older tty-stealing > behaviour which should Just Work. You then have the option of moving over > to /dev/xvc0 once you''ve switched the config over. You''ll need to move to > the more polite /dev/xvc0 route if you want to use the framebuffer console. > > Hope that helps. Does this fix things for you?Why isn''t this generating lots of "I can''t make a domU" posts? Is this not an error everyone experiences? -- Luke St.Clair http://www.lukestclair.org _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
On Thursday 18 October 2007 12:07:43 Luke St.Clair wrote:> However, with extra = "4 xencons=tty", I now get the last line to be > "Setting system clock" after I get "Couldnt get a file descriptor > referring to the console", 6 times. > > So - it seems to be running, but I can''t get a console to fix the > networking. So this approach seems not to work for me.I am experiencing this exact same issue. Host is Ubuntu Server 7.10 (Gutsy), using the Xen 3.1 packages from the distro.> Is this done in menu.lst/grub command line?I am wondering the same thing. I have been trying it several different ways but so far my testing is inconclusive. -Alan Murrell _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Alan Murrell wrote:> On Thursday 18 October 2007 12:07:43 Luke St.Clair wrote: >> However, with extra = "4 xencons=tty", I now get the last line to be >> "Setting system clock" after I get "Couldnt get a file descriptor >> referring to the console", 6 times. >> >> So - it seems to be running, but I can''t get a console to fix the >> networking. So this approach seems not to work for me. > > I am experiencing this exact same issue. Host is Ubuntu Server 7.10 (Gutsy), > using the Xen 3.1 packages from the distro.So you''ve tried xencons=tty in the config file for the VM, and you''re seeing "setting system clock" as the last line in the VM display? _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
On Tuesday 23 October 2007 06:02:16 Luke wrote:> So you''ve tried xencons=tty in the config file for the VM, and you''re > seeing "setting system clock" as the last line in the VM display?That is correct. Here is my VM config: ---------- kernel = "/boot/vmlinuz-2.6.22-14-xen" ramdisk = "/boot/initrd.img-2.6.22-14-xen" root = "/dev/sda1 ro" builder = "linux" extra = "3 xencons=tty" memory = "96" name = "fw01" #uuid = "06ed00fe-1162-4fc4-b5d8-11993ee4a8b9" cpus = "0" vcpus = 1 vif = [ ''bridge=xenbr1'',''bridge=xenbr0'' ] disk = [ ''file:/virtual_machines/domains/fw01/fw01_root_ext3.img,sda1,w'', ''file:/virtual_machines/domains/fw01/fw01_swap.img,sda2,w'' ] on_poweroff = ''destroy'' on_reboot = ''restart'' on_crash = ''restart'' ---------- Without the "xencons=tty" added, the last line I get is: ---------- [32793.502372] EXT3-fs: mounted filesystem with ordered data mode. ---------- and there is a reference in the boot messages previous to it that say: ---------- [32790.414362] xencons_init: Initializing xen vfb; pass xencons=tty to prevent this [32790.415511] Xen virtual console successfully installed as xvc0 ---------- My Xen server is not running X; it''s just a basic, command-line server install. If I did want to use xvc, is there anything additional that I need to configure? I hope this helps in sorting out the issue...? TIA! -Alan Murrell _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
On Thursday 18 October 2007 12:07:43 Luke St.Clair wrote:> However, with extra = "4 xencons=tty", I now get the last line to be > "Setting system clock" after I get "Couldnt get a file descriptor > referring to the console", 6 times.OK, I have made a little progress on this. I mounted my VM image on a loopback and chroot''d to it. I opened /etc/init.d/hwclock.sh and /etc/init.d/hwclockfirst.sh and added the following to the top of each file: exit 0 and it now boots fully into a console. I did run the "update-rc.d" script to remove those two scripts from being run on startup, but it seems that something is still calling them. Anyway, this is at best a workaround, but it works! -Alan Murrell _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Alan Murrell schrieb:> On Thursday 18 October 2007 12:07:43 Luke St.Clair wrote: >> However, with extra = "4 xencons=tty", I now get the last line to be >> "Setting system clock" after I get "Couldnt get a file descriptor >> referring to the console", 6 times. > > OK, I have made a little progress on this. I mounted my VM image on a > loopback and chroot''d to it. I opened /etc/init.d/hwclock.sh > and /etc/init.d/hwclockfirst.sh and added the following to the top of each > file: > > exit 0This issue is caused by an ubuntu domU kernel patch which uses the clock of the hostsystem. I did some similar by chmod 000 /etc/init.d/hwclock* which isn''t a clean solution but works also. A clean solution would be to patch the hwclock scripts to look for e.g. /proc/xenbus . I would file a bugreport, but I assume I''m unable to handle launchpad.net ... Stephan -- Stephan Seitz Senior System Administrator *netz-haut* e.K. multimediale kommunikation zweierweg 22 97074 würzburg fon: +49 931 2876247 fax: +49 931 2876248 web: www.netz-haut.de <http://www.netz-haut.de/> registriergericht: amtsgericht würzburg, hra 5054 _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Stephan Seitz wrote:> Alan Murrell schrieb: >> On Thursday 18 October 2007 12:07:43 Luke St.Clair wrote: >>> However, with extra = "4 xencons=tty", I now get the last line to be >>> "Setting system clock" after I get "Couldnt get a file descriptor >>> referring to the console", 6 times. >> OK, I have made a little progress on this. I mounted my VM image on a >> loopback and chroot''d to it. I opened /etc/init.d/hwclock.sh >> and /etc/init.d/hwclockfirst.sh and added the following to the top of each >> file: >> >> exit 0 > > This issue is caused by an ubuntu domU kernel patch which uses the clock > of the hostsystem. > I did some similar by chmod 000 /etc/init.d/hwclock* > which isn''t a clean solution but works also. > > A clean solution would be to patch the hwclock scripts to look for e.g. > /proc/xenbus . I would file a bugreport, but I assume I''m unable to handle > launchpad.net ... > >I have verified that this fixed my issue as well. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
> >> OK, I have made a little progress on this. I mounted my VM image on a > >> loopback and chroot''d to it. I opened /etc/init.d/hwclock.sh > >> and /etc/init.d/hwclockfirst.sh and added the following to the top of each > >> file: > >> > >> exit 0 > >> I have verified that this fixed my issue as well. >I can second that. I experienced the same issue - gutsy domU''s wouldn''t boot. Disabling the hwclock.sh and hwclockfirst.sh scripts in the domU image fixed the problem. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
> >> OK, I have made a little progress on this. I mounted my VM imageon a> >> loopback and chroot''d to it. I opened /etc/init.d/hwclock.sh > >> and /etc/init.d/hwclockfirst.sh and added the following to the topof each> >> file: > >> > >> exit 0 > >> I have verified that this fixed my issue as well. > >I can second that. I experienced the same issue - gutsy domU''s wouldn''t >boot. >Disabling the hwclock.sh and hwclockfirst.sh scripts in the domU image >fixed the problem.Now some other issues have cropped up. The domu boots up and I can get a console to it, but networking hasn''t started, mount only shows proc and sysfs mounted (but I can still access files on the root filesystem -figure that one out) and my log files are empty: root@wpgsrv5:~# mount proc on /proc type proc (rw) sysfs on /sys type sysfs (rw) root@wpgsrv5:~# ifconfig root@wpgsrv5:~# ls -al /var/log total 192 drwxr-xr-x 5 root root 4096 Nov 14 16:54 . drwxr-xr-x 13 root root 4096 Nov 14 16:49 .. drwxr-xr-x 2 root root 4096 Nov 14 16:55 apt -rw-r----- 1 syslog adm 0 Nov 14 16:54 auth.log -rw-r----- 1 root adm 31 Nov 14 16:51 boot -rw-r--r-- 1 root root 38486 Nov 14 16:54 bootstrap.log -rw-rw-r-- 1 root utmp 0 Nov 14 16:49 btmp -rw-r----- 1 syslog adm 0 Nov 14 16:54 daemon.log -rw-r----- 1 syslog adm 0 Nov 14 16:54 debug -rw-r----- 1 root adm 31 Nov 14 16:51 dmesg -rw-r----- 1 root adm 113627 Nov 14 16:56 dpkg.log -rw-r--r-- 1 root root 2496 Nov 14 16:56 faillog drwxr-xr-x 2 root root 4096 Nov 14 16:51 fsck -rw-r----- 1 syslog adm 0 Nov 14 16:54 kern.log -rw-rw-r-- 1 root utmp 30368 Nov 14 16:56 lastlog -rw-r----- 1 syslog adm 0 Nov 14 16:54 lpr.log -rw-r----- 1 syslog adm 0 Nov 14 16:54 mail.err -rw-r----- 1 syslog adm 0 Nov 14 16:54 mail.info -rw-r----- 1 syslog adm 0 Nov 14 16:54 mail.log -rw-r----- 1 syslog adm 0 Nov 14 16:54 mail.warn -rw-r----- 1 syslog adm 0 Nov 14 16:54 messages drwxr-sr-x 2 news news 4096 Nov 14 16:54 news -rw-r----- 1 syslog adm 0 Nov 14 16:54 syslog -rw-r----- 1 syslog adm 0 Nov 14 16:54 user.log -rw-rw-r-- 1 root utmp 0 Nov 14 16:49 wtmp root@wpgsrv5:~# mount proc on /proc type proc (rw) sysfs on /sys type sysfs (rw) I did find out through Googling that Ubuntu doesn''t like moving /var around, so I checked to see if /var/run and /var/lock exist on the root partition. They do. Can anyone tell me what is going on? Thanks. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users