As I understood you correctly some weeks ago, b129 was set to include xvm 3.4? But apparently it is not: http://pkg.opensolaris.org/dev/en/search.shtml?token=xvm&action=Search Is there again a blocker? Florian
Be that as it may, on my freshly updated system I see this:
root@zeki:~# virsh version
Compiled against library: libvir 0.7.0
Using library: libvir 0.7.0
Using API: Xen 3.0.1
Running hypervisor: Xen 3.4
That said, I also see this:
root@zeki:~# virsh
Welcome to virsh, the virtualization interactive terminal.
Type: ''help'' for help with commands
''quit'' to quit
ld.so.1: virsh: fatal: relocation error: file /usr/lib/64/libreadline.so.5:
symbol tgetent: referenced symbol not found
Killed
cheers,
James
On 12/12/2009, at 8:59 PM, Florian Manschwetus wrote:
> As I understood you correctly some weeks ago, b129 was set to include
> xvm 3.4?
> But apparently it is not:
> http://pkg.opensolaris.org/dev/en/search.shtml?token=xvm&action=Search
>
> Is there again a blocker?
Seems to be working smoothly :-
http://bderzhavets.wordpress.com/2009/12/11/build-sun-xvm-3-4-2-from-source-on-top-osol-1002-128a/
________________________________
From: James Lever <j@jamver.id.au>
To: Florian Manschwetus <florianmanschwetus@gmx.de>
Cc: Xen discuss <xen-discuss@opensolaris.org>
Sent: Sat, December 12, 2009 2:04:34 PM
Subject: Re: [xen-discuss] b129 still no xvm 3.4?
Be that as it may, on my freshly updated system I see this:
root@zeki:~# virsh version
Compiled against library: libvir 0.7.0
Using library: libvir 0.7.0
Using API: Xen 3.0.1
Running hypervisor: Xen 3.4
That said, I also see this:
root@zeki:~# virsh
Welcome to virsh, the virtualization interactive terminal.
Type: ''help'' for help with commands
''quit'' to quit
ld.so.1: virsh: fatal: relocation error: file /usr/lib/64/libreadline.so.5:
symbol tgetent: referenced symbol not found
Killed
cheers,
James
On 12/12/2009, at 8:59 PM, Florian Manschwetus wrote:
> As I understood you correctly some weeks ago, b129 was set to include
> xvm 3.4?
> But apparently it is not:
> http://pkg.opensolaris.org/dev/en/search.shtml?token=xvm&action=Search
>
> Is there again a blocker?
_______________________________________________
xen-discuss mailing list
xen-discuss@opensolaris.org
James Lever wrote:> Be that as it may, on my freshly updated system I see this: > > root@zeki:~# virsh version > Compiled against library: libvir 0.7.0 > Using library: libvir 0.7.0 > Using API: Xen 3.0.1 > Running hypervisor: Xen 3.4 > > That said, I also see this: > > root@zeki:~# virsh > Welcome to virsh, the virtualization interactive terminal. > > Type: ''help'' for help with commands > ''quit'' to quit > > ld.so.1: virsh: fatal: relocation error: file /usr/lib/64/libreadline.so.5: symbol tgetent: referenced symbol not found > Killed >3.4 is in build 129. The readline issue is known: the libvirt package lacks a dependency on readline. To fix this, you can simply do: pkg install SUNWgnu-readline ..and you should be ok. - Frank
> James Lever wrote: > 3.4 is in build 129. The readline issue is known: the > libvirt package > lacks a dependency on readline. To fix this, you can > simply do: > > pkg install SUNWgnu-readline > > ..and you should be ok. > > - FrankI already had readline installed, however i tried reinstalling and I''m still unable to enter an virsh interactive shell. :\ -- This message posted from opensolaris.org
I an getting the issue similar "svcadm enable milestone/xvm" with xvm 3.3.2. At 5-th or 6-th boot up gnome starts extremely slow and finally reports "Error Window - Clock Applet cannot be loaded", when i choose "delete" or "save" pretty slow gnome environment comes up. Next "reboot -p" causes a bunch of messages "No swap space available" scrolling through the screen until maintenance prompt. I was able to reproduce the issue on 2 different boxes. After first 3 - 4 cold or warm bootups everything works fine , including managing VMs via virsh. -- This message posted from opensolaris.org
I removed SUNWgnu-readline and all of its dependencies -- after re-installing, the issue persists. -- This message posted from opensolaris.org
Dan Reiland wrote:> I removed SUNWgnu-readline and all of its> dependencies -- after re-installing, the issue persists. This is the following bug (fixed in b130) 6905338 virsh throwing error message in snv_129 and osol_129 The official build machine had an old version of readline which had a dependency on curses. You can still run virsh, just not interactively... i.e. all the virsh commands still work. e.g. virsh list virsh start mydomain. There are a couple of workarounds... You can put the old readline library ( b125 or older) which still had the dependency on curses on your machine.. You can grab the b129 xvm source and rebuild on a >b126 machine and update the SUNWlibvirt package or just the virsh bin. MRJ
Boris Derzhavets wrote:> I an getting the issue similar "svcadm enable milestone/xvm" with xvm 3.3.2. At 5-th or 6-th boot up gnome starts extremely slow and finally reports "Error Window - Clock Applet cannot be loaded", > when i choose "delete" or "save" pretty slow gnome environment comes up. > Next "reboot -p" causes a bunch of messages "No swap space available" scrolling through the > screen until maintenance prompt. I was able to reproduce the issue on 2 different boxes. > After first 3 - 4 cold or warm bootups everything works fine , including managing VMs via virsh.What build are you on? There are a bunch of milestone/xvm bug fixes in b129. In b129 I can svcadm enable milestone/xvm when booted into metal then reboot and everything works fine for me. MRJ
I am on 128a. Build ( described above) xVM 3.4.2 packages via mercurial has been done on on 128a. I am not quite sure that procedure itself was correct. Have all packages required for 3.4.2 build been installed or no ? Date of xen changeset was 12/11/09. ________________________________ From: Mark Johnson <Mark.Johnson@Sun.COM> To: Boris Derzhavets <bderzhavets@yahoo.com> Cc: xen-discuss@opensolaris.org Sent: Mon, December 14, 2009 1:09:48 AM Subject: Re: [xen-discuss] b129 still no xvm 3.4? Boris Derzhavets wrote:> I an getting the issue similar "svcadm enable milestone/xvm" with xvm 3.3.2. At 5-th or 6-th boot up gnome starts extremely slow and finally reports "Error Window - Clock Applet cannot be loaded", > when i choose "delete" or "save" pretty slow gnome environment comes up. > Next "reboot -p" causes a bunch of messages "No swap space available" scrolling through the > screen until maintenance prompt. I was able to reproduce the issue on 2 different boxes. > After first 3 - 4 cold or warm bootups everything works fine , including managing VMs via virsh.What build are you on? There are a bunch of milestone/xvm bug fixes in b129. In b129 I can svcadm enable milestone/xvm when booted into metal then reboot and everything works fine for me. MRJ
I just installed osol_129 with native xVM 3.4.2 ( 11/24/09). Yes, it reboots fine. However :- # virt-install -n VMF12 -r 1024 -p --vnc --os-type=linux --os-variant=fedora11 \ -f /dev/zvol/dsk/tank1/disk1 -l http://192.168.1.33/f12 --debug hangs attempting to bring up VNC console, same virt-install runs fine on xVM 3.4.2 (12/11/09) been built on osol_128a via mercurial source. ________________________________ From: Mark Johnson <Mark.Johnson@Sun.COM> To: Boris Derzhavets <bderzhavets@yahoo.com> Cc: xen-discuss@opensolaris.org Sent: Mon, December 14, 2009 3:09:48 AM Subject: Re: [xen-discuss] b129 still no xvm 3.4? Boris Derzhavets wrote:> I an getting the issue similar "svcadm enable milestone/xvm" with xvm 3.3.2. At 5-th or 6-th boot up gnome starts extremely slow and finally reports "Error Window - Clock Applet cannot be loaded", > when i choose "delete" or "save" pretty slow gnome environment comes up. > Next "reboot -p" causes a bunch of messages "No swap space available" scrolling through the > screen until maintenance prompt. I was able to reproduce the issue on 2 different boxes. > After first 3 - 4 cold or warm bootups everything works fine , including managing VMs via virsh.What build are you on? There are a bunch of milestone/xvm bug fixes in b129. In b129 I can svcadm enable milestone/xvm when booted into metal then reboot and everything works fine for me. MRJ
I just installed osol_129 with native xVM 3.4.2 ( 11/24/09). Yes, it reboots fine. However :- # virt-install -n VMF12 -r 1024 -p --vnc --os-type=linux --os-variant=fedora11 \ -f /dev/zvol/dsk/tank1/disk1 -l http://192.168.1.33/f12 --debug hangs attempting to bring up VNC console, same virt-install runs fine on xVM 3.4.2 (12/11/09) been built on osol_128a via mercurial source. At the same time on osol_129 with native xVM 3.4.2 ( 11/24/09) : virt-install -n f12p -r 1024 -p --nographics --os-type=linux --os-variant=fedora11 \ -f /dev/zvol/dsk/tank2/disk1 -l http://192.168.1.33/f12 --debug starts with no problems. -- This message posted from opensolaris.org
Ok, I updated one machine, with some problems depending on jumping from 0906 to b129, never mind, yes it provides xvm 3.4, so the package information seems to be wrong or missleading. http://pkg.opensolaris.org/dev/info/0/SUNWxvm%403.3.2%2C5.11-0.129%3A20091205T133022Z (see version tag) florian On 14.12.2009 01:07, Mark Johnson wrote:> > > Dan Reiland wrote: >> I removed SUNWgnu-readline and all of its >> dependencies -- after re-installing, the issue persists. > > This is the following bug (fixed in b130) > 6905338 virsh throwing error message in snv_129 and osol_129 > > The official build machine had an old version of readline > which had a dependency on curses. > > You can still run virsh, just not interactively... i.e. > all the virsh commands still work. > e.g. virsh list > virsh start mydomain. > > There are a couple of workarounds... > > You can put the old readline library ( b125 or older) > which still had the dependency on curses on your machine.. > > You can grab the b129 xvm source and rebuild on a >> b126 machine and update the SUNWlibvirt package or > just the virsh bin. > > > MRJ > > _______________________________________________ > xen-discuss mailing list > xen-discuss@opensolaris.org >
root@opensolaris129:/var/log/xen# cat qemu-dm-F12PV.log domid: 14 Using xvda for guest''s hda Watching /local/domain/0/device-model/14/logdirty/next-active Watching /local/domain/0/device-model/14/command Warning: vlan 0 is not connected to host network char device redirected to /dev/pts/2 xs_read(): vncpasswd get error. /vm/fb0b2f87-65f4-a5a3-c814-282f95d68e7e/vncpasswd -- This message posted from opensolaris.org
Attempt to start F12 PV DomU via virt-manager ( http - source, zvol - target) appears to be able to start domain creation, but console supposed to be connected to guest is empty again. Seems as a bug in osol_129. -- This message posted from opensolaris.org
Boris Derzhavets wrote:> I just installed osol_129 with native xVM 3.4.2 ( 11/24/09). Yes, it reboots fine. > However :- > > # virt-install -n VMF12 -r 1024 -p --vnc --os-type=linux --os-variant=fedora11 \ > -f /dev/zvol/dsk/tank1/disk1 -l http://192.168.1.33/f12 --debug > > hangs attempting to bring up VNC console, same virt-install runs fine on xVM 3.4.2 (12/11/09) > been built on osol_128a via mercurial source. > > At the same time on osol_129 with native xVM 3.4.2 ( 11/24/09) : > > virt-install -n f12p -r 1024 -p --nographics --os-type=linux --os-variant=fedora11 \ > -f /dev/zvol/dsk/tank2/disk1 -l http://192.168.1.33/f12 --debug > > starts with no problems.Try doing a xm console or virsh console to the guest. I have a report that someone saw something similar and the vnc window popped up after they connected to the serial console. MRJ
Florian Manschwetus wrote:> Ok, I updated one machine, with some problems depending on jumping from > 0906 to b129, never mind, yes it provides xvm 3.4, so the package > information seems to be wrong or missleading.Thanks, the package name will be fixed in osol 130. MRJ> http://pkg.opensolaris.org/dev/info/0/SUNWxvm%403.3.2%2C5.11-0.129%3A20091205T133022Z > > (see version tag) > > florian > > On 14.12.2009 01:07, Mark Johnson wrote: >> >> Dan Reiland wrote: >>> I removed SUNWgnu-readline and all of its >>> dependencies -- after re-installing, the issue persists. >> This is the following bug (fixed in b130) >> 6905338 virsh throwing error message in snv_129 and osol_129 >> >> The official build machine had an old version of readline >> which had a dependency on curses. >> >> You can still run virsh, just not interactively... i.e. >> all the virsh commands still work. >> e.g. virsh list >> virsh start mydomain. >> >> There are a couple of workarounds... >> >> You can put the old readline library ( b125 or older) >> which still had the dependency on curses on your machine.. >> >> You can grab the b129 xvm source and rebuild on a >>> b126 machine and update the SUNWlibvirt package or >> just the virsh bin. >> >> >> MRJ >> >> _______________________________________________ >> xen-discuss mailing list >> xen-discuss@opensolaris.org >> > >
Yes, # virsh console <DomName> brings vnc-console up Thanks. Boris. --------------------------------------------------------------------------------------------------------- From: Mark Johnson <Mark.Johnson@Sun.COM> To: Boris Derzhavets <bderzhavets@yahoo.com> Cc: xen-discuss@opensolaris.org Sent: Mon, December 14, 2009 10:20:24 PM Subject: Re: [xen-discuss] b129 still no xvm 3.4? Boris Derzhavets wrote:> I just installed osol_129 with native xVM 3.4.2 ( 11/24/09). Yes, it reboots fine. > However :- > > # virt-install -n VMF12 -r 1024 -p --vnc --os-type=linux --os-variant=fedora11 \ > -f /dev/zvol/dsk/tank1/disk1 -l http://192.168.1.33/f12 --debug > > hangs attempting to bring up VNC console, same virt-install runs fine on xVM 3.4.2 (12/11/09) > been built on osol_128a via mercurial source. > > At the same time on osol_129 with native xVM 3.4.2 ( 11/24/09) : > > virt-install -n f12p -r 1024 -p --nographics --os-type=linux --os-variant=fedora11 \ > -f /dev/zvol/dsk/tank2/disk1 -l http://192.168.1.33/f12 --debug > > starts with no problems.Try doing a xm console or virsh console to the guest. I have a report that someone saw something similar and the vnc window popped up after they connected to the serial console. MRJ