Re: [opensuse-virtual] Leap 42.2 + Xen: Missing Independent Wallclock
On 11/29/2016 11:02 PM, Rieß Norman wrote: Am 29.11.2016 um 16:40 schrieb Jim Fehlig : Rieß Norman wrote: Am 29.11.2016 um 08:48 schrieb Olaf Hering : On Tue, Nov 29, Rieß Norman wrote: we started using independent wallclock because the little timejump while live migrating was sufficient to disrupt some time dependent services. All our hosts (and VMs) use NTP, but the drift seems to be enough. Setting independent wallclock solved this behaviour. Are you saying that even with correctly running ntpd in both dom0s the migrated domU will see a timejump? A small jump forward is expected because the domU goes through a suspend/resume cycle while the last pages of dirty memory get transfered. I would be surprised if independent wallclock can change anything in this regard. Olaf Well, yes. This was the case several years ago. As the problem disappeared with the use of independent wallclock, we did not investigate in later releases. To summarize the takeaway from this thread: The independent_wallclock control does not exist in the pvops kernel. If absence of this control causes problems, please enter a bug with the details. Regards, Jim You should update your documentation then. Yep. If you care to contribute to openSUSE, patches welcome :-). Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] Leap 42.2 + Xen: Missing Independent Wallclock
Rieß Norman wrote: >> Am 29.11.2016 um 08:48 schrieb Olaf Hering : >> >> On Tue, Nov 29, Rieß Norman wrote: >> >>> we started using independent wallclock because the little timejump >>> while live migrating was sufficient to disrupt some time dependent >>> services. All our hosts (and VMs) use NTP, but the drift seems to be >>> enough. Setting independent wallclock solved this behaviour. >> Are you saying that even with correctly running ntpd in both dom0s the >> migrated domU will see a timejump? A small jump forward is expected >> because the domU goes through a suspend/resume cycle while the last >> pages of dirty memory get transfered. I would be surprised if >> independent wallclock can change anything in this regard. >> >> Olaf > > Well, yes. This was the case several years ago. As the problem disappeared > with the use of independent wallclock, we did not investigate in later > releases. To summarize the takeaway from this thread: The independent_wallclock control does not exist in the pvops kernel. If absence of this control causes problems, please enter a bug with the details. Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] KVM + Kimchi on Leap 42.1
On 07/08/2016 02:43 PM, Ciro Iriarte wrote: > 2016-07-08 13:06 GMT-04:00 Jim Fehlig : >> On 07/08/2016 10:59 AM, Ciro Iriarte wrote: >>> Hi!, I just did a fresh install of Leap 42.1 + updates + >>> Virtualization repo, and trying to start the service I get a timeout >>> on the libvirtd.service dependency, all I see is a "unable to notify >>> systemd" message. >>> >>> Anybody has seen this?, any suggestion?. >> Yes, see http://bugzilla.opensuse.org/show_bug.cgi?id=987668 >> >> Regards, >> Jim > Would Leap 42.2 Alpha work as a work around?. Yes, it should. Leap 42.2 currently contains systemd v228, which is the same as the latest SLE12 SP2 beta where NOTIFY_SOCKET works just fine. Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] KVM + Kimchi on Leap 42.1
On 07/08/2016 10:59 AM, Ciro Iriarte wrote: > Hi!, I just did a fresh install of Leap 42.1 + updates + > Virtualization repo, and trying to start the service I get a timeout > on the libvirtd.service dependency, all I see is a "unable to notify > systemd" message. > > Anybody has seen this?, any suggestion?. Yes, see http://bugzilla.opensuse.org/show_bug.cgi?id=987668 Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] how to update the config for a xen domu
On 02/29/2016 01:48 AM, Per Jessen wrote: > Olaf Hering wrote: > >> On Fri, Feb 26, Per Jessen wrote: >> >>> a xen domu is created from config-file on hostA, then migrated to >>> hostB. On hostB, a new disk is added with "xl block-attach", works >>> very well. However "xl list --long " doesn't reflect this >>> change. "xl config-update" seems to be a way to edit config files, >>> but once the domU has been migrated to hostB, the original file isn't >>> immediately available - what do I need to get the new block device >>> reflected in the config that goes with the domU ? >> I think that requires a newer libxl version which tracks the config in >> a separate JSON file. Until then the domU.cfg has to be adjusted along >> with the 'xl block-attach'. > Thanks Olaf, I suspected that too. > > Just wondering what other people do when you have multiple xen hosts and > you migrate domUs between them - do you keep the configs on NFS, shared > by all hosts? If you don't mind another layer of tooling, you can use libvirt to manage your xen domains. It will take care of managing state/config for both transient and persistent domains. It will also transfer the current state when migrating the domain, optionally persisting the state/config on the target host when migration is invoked with the --persistent flag (virsh migrate ... --persistent). Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] pvgrub2-boot of Opensuse Xen DomU -- possible? documentation? example?
On 10/30/2015 02:55 PM, PGNet Dev wrote: > On 10/30/2015 01:38 PM, Charles Arnold wrote: >> Here is a generic (but working) example of a config file. >> What I am not using in this example is the 'root' option. >> >> name="os13.2pv" >> description="None" >> uuid="09e84b64-5b3d-bb48-7fe3-488e7058878d" >> memory=1024 >> maxmem=1024 >> vcpus=2 >> on_poweroff="destroy" >> on_reboot="restart" >> on_crash="destroy" >> localtime=0 >> keymap="en-us" >> builder="linux" >> kernel="/usr/lib/grub2/x86_64-xen/grub.xen" >> disk=[ '/dev/dm-0,raw,xvda,w,backendtype=phy', ] >> vif=[ 'mac=00:16:3e:0d:cb:82,bridge=br0', ] >> vfb=['type=vnc,vncunused=1'] > > > I note your use of format > > disk=[ '/dev/dm-0,raw,xvda,w,backendtype=phy', ] > > Noting that my usage boots just fine using pygrub, > > disk = [ 'phy:/dev/VG0/testBOOT,xvda,w', > 'phy:/dev/VG0/testROOT,xvdc,w', > ] > > is a change to 'your' format *required* here? I haven't followed this discussion, but the the answer to the general question you ask here is no. The first format is the new xl disk config format, described in $xen-src/docs/misc/xl-disk-configuration.txt and xl.cfg(5). The second format is the old xmdomain.cfg(5) format. But the old format should still be supported. With the exception of embedded python, a Xen community goal was to support existing domain config when moving from xm/xend to xl/libxl. > > > Also, instead of a root-vol spec'n, > > root = '(xen/xvda,msdos1)' > > you appear to use a UUID spec, > > uuid="09e84b64-5b3d-bb48-7fe3-488e7058878d" In this context, uuid is simply domain metadata, similar to name, id, description, etc. Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] Where to set "/xen.gz " in grub2 config?
lynd...@your-mail.com wrote: > On Wed, Mar 25, 2015, at 08:51 AM, Jim Fehlig wrote: > >> lynd...@your-mail.com wrote: >> >>> I just installed openSUSE 13.2 with Xen & kernel-xen. >>> >>> I need to set up the /xen.gz ... ... in grub. >>> >>> In grub-legacy that's easy. Just modify the /boot/grub/menu.lst >>> >>> kernel /xen.gz dom0_mem=1024M,max:1024M dom0_max_vcpus=2 >>> dom0_vcpus_pin=true ... >>> >>> The system installed grub2 as the default. In grub2 it's not clear to me >>> how to set those options. >>> >>> >> Add the options to GRUB_CMDLINE_XEN in /etc/default/grub and recreate >> /boot/grub2/grub.cfg with 'grub2-mkconfig -o /boot/grub2/grub.cfg'. >> > > Thanks, GRUB_CMDLINE_XEN= takes care of the xen.gz line. > > If I have multikernel enabled and have both a -xen and -default kernel > available, the > > module /vmlinuz-3.19.2-2.g6c3d33f-xen ... > > line looks like it's assembled from a combination of > > GRUB_CMDLINE_LINUX= & GRUB_CMDLINE_LINUX_DEFAULT= > Right. > I need different options on the Xen instance's /vmlinuz line. For example a > different elevator= & clocksource=. > Hmm, not sure if that is currently possible, but I'm not a grub2 expert. Perhaps others on the list have suggestions. Otherwise might be a feature request for grub2. Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] Where to set "/xen.gz " in grub2 config?
lynd...@your-mail.com wrote: > I just installed openSUSE 13.2 with Xen & kernel-xen. > > I need to set up the /xen.gz ... ... in grub. > > In grub-legacy that's easy. Just modify the /boot/grub/menu.lst > > kernel /xen.gz dom0_mem=1024M,max:1024M dom0_max_vcpus=2 > dom0_vcpus_pin=true ... > > The system installed grub2 as the default. In grub2 it's not clear to me how > to set those options. > Add the options to GRUB_CMDLINE_XEN in /etc/default/grub and recreate /boot/grub2/grub.cfg with 'grub2-mkconfig -o /boot/grub2/grub.cfg'. Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] Upgrading Opensuse 13.2 kernel in a Xen Guest breaks networking and access to it. Downgrading it fixes it. What's wrong?
Tony Su wrote: > Thx all, > Yes, it does look then like kernel-xen is required for both Dom0 and > DomU for the kernels openSUSE/SUSE currently provides. > It is not required and shouldn't be used in Xen HVM domU. HVM domUs, like KVM domains, use kernel-default + pv drivers. Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] Upgrading Opensuse 13.2 kernel in a Xen Guest breaks networking and access to it. Downgrading it fixes it. What's wrong?
Tony Su wrote: > Additional information selecting kernels for use in Xen (Dom0 or DomU) > http://wiki.xen.org/wiki/Xen_Kernel_Feature_Matrix > That is upstream documentation. Distros may be configured differently. > Unless I am misunderstanding the information on this official source > - All modern implementations of Xen features for DomU is now > implemented through the pvops framework. > - The pvops framework is part of the mainline Linux kernel. > Right. > This suggests that a special kernel-xen is not needed for DomU and > supports all current documented features. > The pvops kernel is missing features wrt to the xenified kernel. SUSE is actively working on closing the feature gap and moving to pvops, but currently we still use the xenified kernel (kernel-xen). > So, maybe kernel-xen used for DomU would be just another choice to > avoid possible issues in another kernel choice but not necessarily > recommended? > It's not only recommended, but required in a PV VM. kernel-default, which currently doesn't have pvops enabled, wont boot in a PV VM. Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] Upgrading Opensuse 13.2 kernel in a Xen Guest breaks networking and access to it. Downgrading it fixes it. What's wrong?
Tony Su wrote: > > User error. I posted answer in openSUSE technical help forums. > I assume you are talking about this answer: "The xen kernel is installed in the HostOS (if needed). You do install the xen kernel into a Guest (unless you're trying to deploy multiple layers of virtualization which is highly inadvisable). A "normal" kernel you'd normally install in a non-virtual environment should be used in your Guest, ie desktop, default, vanilla, etc." But that is not true. kernel-xen is most certainly used in a PV guest. And in my case, the latest 13.2 kernel-xen works fine, but the latest from the Kernel:stable project does not. Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] Upgrading Opensuse 13.2 kernel in a Xen Guest breaks networking and access to it. Downgrading it fixes it. What's wrong?
ale...@vfemail.net wrote: > On 2015-03-11 10:16, Jan Beulich wrote: >> supply the needed information and help >> with possible further debugging steps. > > I can try, but I read in that bug where it says to give 'dmesg' and > 'xl dmesg' logs. I was able to see similar networking problems with the latest kernel-xen from http://download.opensuse.org/repositories/Kernel:/stable/standard and have reopened the bug, providing some initial info. Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] adding a xen guest takes down interface of existing guest??
Per Jessen wrote: > Per Jessen wrote: > > >> My xen host is openSUSE 13.2. >> >> I have one guest running currently. I created a second guest to >> install 13.2 and ran "xl create". >> >> This appears to have brought down the interface (vif10) for the >> existing guest ?? >> >> 9: vif10: mtu 1500 qdisc pfifo_fast state DOWN >> group default qlen 512 >> link/ether fe:ff:ff:ff:ff:ff brd ff:ff:ff:ff:ff:ff >> >> > > I'm not certain but this seems to be related to me having chosen "vif10" > and "vif11" as the interface names. On a second try, when I omitted > the interface name for the new guest, the new interface was > named "vif11.0" and did not interfere with vif10. I don't see your problem on an updated 13.2 Xen host. I started two PV domains with 'vifname=vif10' and 'vifname=vif11' in their respective 'vif = [...]' config. The vif device of the first domain (vif10) remains up and continues to work when starting the second domain containing vif11. Are your domains PV or HVM? Can you provide the vif config for the domains? Does /var/log/messages contain any hint about who (or what) is bringing down the vif? Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] handling of up/down events from brdige interfaces for KVM and Xen
Olaf Hering wrote:
> I'm about to remove /etc/sysconfig/network/if-up.d/xen which is supposed
> to collect vif* and tap* interfaces connected to a bridge interface.
> Right now it runs only when xend is running, and today wicked will not call
> such scripts anyway.
>
Ah, interesting. I didn't know those would no longer be called.
> I wonder how KVM deals with disappearing bridges?
It doesn't.
> In the end any
> replacement script can very well handle KVM, Xen and whatever host type
> is in use. So a Xen specific tool is not required.
>
Agreed. I now recall brief discussions with Marius around this topic.
In fact, I even created a hack in libvirt (attached) to record the
adding/removing of vif/tap devices from the bridge. I never got around
to thoroughly testing this hack to gain enough confidence to add it to
our libvirt package.
Jim
commit 9a06bafecb099a7879cb92fd6026db9ec594a643
Author: Jim Fehlig
Date: Wed Jun 19 10:16:04 2013 -0600
Register ports connected to a bridge
When connecting a port to a bridge, register the connection
in a well-known location (/var/lib/run/libvirt/port_registration/$portname).
Remove the registration when the port is removed from the bridge.
Hypervisor drivers that use virNetDevBridge{Add,Remove}Port() will have
their virtual interfaces which connect to bridges registered, allowing
reconnection of those interfaces upon bridge restart.
diff --git a/src/util/virnetdevbridge.c b/src/util/virnetdevbridge.c
index 130829c..9956b50 100644
--- a/src/util/virnetdevbridge.c
+++ b/src/util/virnetdevbridge.c
@@ -34,6 +34,9 @@
#include
#include
#include
+#include
+#include
+#include
#ifdef __linux__
# include
@@ -296,6 +299,8 @@ int virNetDevBridgeAddPort(const char *brname,
int fd = -1;
int ret = -1;
struct ifreq ifr;
+char *reg_path;
+int reg_fd;
if ((fd = virNetDevSetupControl(brname, &ifr)) < 0)
return -1;
@@ -312,6 +317,20 @@ int virNetDevBridgeAddPort(const char *brname,
goto cleanup;
}
+/*
+ * Register the port in a well known location so it can be reconnected
+ * when the hosting bridge is restarted.
+ */
+if (virAsprintf(®_path,
+"/var/run/libvirt/port_registrations/%s",
+ifname) > 0) {
+if ((reg_fd = open(reg_path, O_CREAT)) > 0) {
+ignore_value(safewrite(reg_fd, brname, strlen(brname)));
+VIR_FORCE_CLOSE(reg_fd);
+}
+VIR_FREE(reg_path);
+}
+
ret = 0;
cleanup:
VIR_FORCE_CLOSE(fd);
@@ -343,6 +362,17 @@ int virNetDevBridgeRemovePort(const char *brname,
int fd = -1;
int ret = -1;
struct ifreq ifr;
+char *reg_path;
+
+/*
+ * Remove any port registration.
+ */
+if (virAsprintf(®_path,
+"/var/run/libvirt/port_registrations/%s",
+ifname) > 0) {
+ignore_value(unlink(reg_path));
+VIR_FREE(reg_path);
+}
if ((fd = virNetDevSetupControl(brname, &ifr)) < 0)
return -1;
Re: [opensuse-virtual] libvirtd, qemu+ssh and polkit: internal error: received hangup / error event on socket?
On 03/24/2014 02:49 PM, Johannes Kastl wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Hi everyone, maybe this is an old issue, but I am just starting to set up my machines with libvirt and qemu-kvm, and tried using a remote connection. I already changed the following lines in libvirtd.conf: unix_sock_group = "libvirt" unix_sock_rw_perms = "0770" auth_unix_rw = "none" I made my user a member of said group 'libvirt'. And I restarted both libvirtd and polkit. I can connect as local user via virsh -c qemu:///sytem without a password prompt. I can also ssh into the remote host with my ssh key without problems. Trying from another machine (say hostname HOSTX, username testuser) with virsh -c qemu+ssh://testusers@HOSTX/system always results in error: failed to connect to the hypervisor error: internal error: received hangup / error event on socket Unfortunately, you can't use polkit remotely. I think you are suffering the same problem I wasted time investigating several years ago http://www.redhat.com/archives/libvir-list/2010-November/msg00820.html Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] Xl usb-attach in opensuse 13.1
Jeroen Groenewegen van der Weyden wrote: > Antoine Ginies schreef op 12-2-2014 10:57: >> Jeroen Groenewegen van der Weyden: >>> Hello, >>> >>> I updated my machines recently from opensuse 12.3 to 13.1. >>> Which made us switch from xm to xl. However there is no command >>> available for xl to attach on usb device to a running domain. >> you can use: >> virsh attach-device > Well virsh does not seem to function properly either, the output of > virsh list is empty. Antoine already gave a hint, but thought I'd point out a blog post about libvirt's use of the new Xen toolstack. It mentions some differences between the old libvirt Xen driver and the new libxl one http://jfehlig.wordpress.com/2014/01/05/libvirt-support-for-xens-new-libxenlight-toolstack/ Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] Xl usb-attach in opensuse 13.1
Antoine Ginies wrote:
> Jeroen Groenewegen van der Weyden:
>
>> Hello,
>>
>> I updated my machines recently from opensuse 12.3 to 13.1.
>> Which made us switch from xm to xl. However there is no command
>> available for xl to attach on usb device to a running domain.
>>
>
> you can use:
> virsh attach-device
>
> and:
> virsh detach-device
>
> Yet there is no xl command to do that.
>
There are no xl commands because there is no support in libxl. libvirt
also uses libxl, so this wont work in libvirt either.
scsi-{attach,detach,list} is also missing from the libxl stack.
Regards,
Jim
--
To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org
To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] XEN on a 13.1 system
Hans Witvliet wrote:
> On Mon, 2014-01-20 at 19:29 -0700, Jim Fehlig wrote:
>
>> Hans Witvliet wrote:
>>
>>> Yes, i forgotten that xm/xend is depreceated.
>>> Heard that a couple of times at several fosdems from Ian,
>>> and when you mention "xm" at the xen-lists you are considered a relic
>>> from a distant past
>>>
>>>
>>> How ever:
>>> When you do a fresh instalation of the xen-stuff in 13.1, and you try to
>>> create a VM like one is used to do in previous versions, i still got a
>>> warning that i had to check because "xend was not running"
>>>
>>>
>> Charles, is there a check in {vm,virt}-install for xend running? I
>> vaguely recall something along those lines, but suppose it was fixed
>> long ago.
>>
>> WRT libvirt, If xend is disabled and libvirtd is restarted, it will use
>> the libxl driver, and never fail or warn due to "xend not running".
>>
>>
>>> btw the message i got when trying to build it from yast was:
>>> "Error Failed to start the vm, Details:
>>> Error, device0 (VIF) could not be connected.
>>> Hotplug scripts are not working"
>>>
>>>
>> I think this is due to a mis-configured toolstack. We've also fixed a
>> number of bugs since 13.1 was released. I assume you have all the
>> latest updates? You can also check for relevant fixes that are queued
>> for the next update in our 13.1 maintenance project
>>
>> https://build.opensuse.org/project/show/Virtualization:openSUSE13.1
>>
>> Regards,
>> Jim
>>
>>
> Hi Jim,
>
>
> yes lastest patches were included (during installation).
>
> Is it worthwhile trying to add the stuff in:
> repositories/Virtualization:/openSUSE13.1/openSUSE_13.1/x86_64/
> the libvirt rpm's are from 21-ste of januari ?
>
Yes, there are a few fixes for the libxl driver queued in that repo for
the *next* 13.1 update. I.e., fixes that have yet to hit the 13.1
update repos.
Regards,
Jim
--
To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org
To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] XEN on a 13.1 system
Charles Arnold wrote:
>>>> On 1/20/2014 at 07:29 PM, in message <52dddba2.5000...@suse.com>, Jim
>>>> Fehlig
>>>>
> wrote:
>
>> Hans Witvliet wrote:
>>
>>> Yes, i forgotten that xm/xend is depreceated.
>>> Heard that a couple of times at several fosdems from Ian,
>>> and when you mention "xm" at the xen-lists you are considered a relic
>>> from a distant past
>>>
>>>
>>> How ever:
>>> When you do a fresh instalation of the xen-stuff in 13.1, and you try to
>>> create a VM like one is used to do in previous versions, i still got a
>>> warning that i had to check because "xend was not running"
>>>
>>>
>> Charles, is there a check in {vm,virt}-install for xend running? I
>> vaguely recall something along those lines, but suppose it was fixed
>> long ago.
>>
>
> Yes. vm-install does check for a running xend. This is how it knows
> whether to use 'xm' to start VMs or if xend is not running it uses a
> libvirt connection to start VMs. The only time vm-install uses 'xl' for
> starting VMs is if xend is not running and it fails to connect to libvirt.
> Failure to connect could mean libvirtd was not running or some other
> error occurred when connecting to libvirtd.
>
BTW, I don't think we should support xl in vm-install. If vm-install
cannot connect to libvirt, then it should just fail IMO. That seems to
be the behavior on KVM
# systemctl stop libvirtd.service
# vm-install
libvirt: XML-RPC error : Failed to connect socket to
'/var/run/libvirt/libvirt-sock': No such file or directory
Otherwise users might be able to create VM's, which they are then unable
to see or manage with virt-manager.
Jim
--
To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org
To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] XEN on a 13.1 system
Charles Arnold wrote:
>>>> On 1/20/2014 at 07:29 PM, in message <52dddba2.5000...@suse.com>, Jim
>>>> Fehlig
>>>>
> wrote:
>
>> Hans Witvliet wrote:
>>
>>> Yes, i forgotten that xm/xend is depreceated.
>>> Heard that a couple of times at several fosdems from Ian,
>>> and when you mention "xm" at the xen-lists you are considered a relic
>>> from a distant past
>>>
>>>
>>> How ever:
>>> When you do a fresh instalation of the xen-stuff in 13.1, and you try to
>>> create a VM like one is used to do in previous versions, i still got a
>>> warning that i had to check because "xend was not running"
>>>
>>>
>> Charles, is there a check in {vm,virt}-install for xend running? I
>> vaguely recall something along those lines, but suppose it was fixed
>> long ago.
>>
>
> Yes. vm-install does check for a running xend. This is how it knows
> whether to use 'xm' to start VMs or if xend is not running it uses a
> libvirt connection to start VMs. The only time vm-install uses 'xl' for
> starting VMs is if xend is not running and it fails to connect to libvirt.
> Failure to connect could mean libvirtd was not running or some other
> error occurred when connecting to libvirtd.
>
> For virt-install, it only uses libvirt for starting VMs. It will try to
> default
> to a xen connection if '/var/lib/xend' and '/proc/xen' exist otherwise it
>
It shouldn't look for /var/lib/xend, since that directory wont exist on
a system without xend.
Jim
> looks to see if qemu/kvm is present and connects to that. With
> virt-manager, and the command line virt-install you can specify the
> connection you want and that is what virt-install will use assuming the
> appropriate libvirt driver was loaded.
>
> - Charles
>
>
--
To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org
To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] XEN on a 13.1 system
Hans Witvliet wrote:
> Yes, i forgotten that xm/xend is depreceated.
> Heard that a couple of times at several fosdems from Ian,
> and when you mention "xm" at the xen-lists you are considered a relic
> from a distant past
>
>
> How ever:
> When you do a fresh instalation of the xen-stuff in 13.1, and you try to
> create a VM like one is used to do in previous versions, i still got a
> warning that i had to check because "xend was not running"
>
Charles, is there a check in {vm,virt}-install for xend running? I
vaguely recall something along those lines, but suppose it was fixed
long ago.
WRT libvirt, If xend is disabled and libvirtd is restarted, it will use
the libxl driver, and never fail or warn due to "xend not running".
> btw the message i got when trying to build it from yast was:
> "Error Failed to start the vm, Details:
> Error, device0 (VIF) could not be connected.
> Hotplug scripts are not working"
>
I think this is due to a mis-configured toolstack. We've also fixed a
number of bugs since 13.1 was released. I assume you have all the
latest updates? You can also check for relevant fixes that are queued
for the next update in our 13.1 maintenance project
https://build.opensuse.org/project/show/Virtualization:openSUSE13.1
Regards,
Jim
--
To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org
To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] XEN on a 13.1 system
Hans Witvliet wrote: > Hi all, > > Are there any known issues with 13.1 as a DOM0 ? > > @work, i installed+updated a new system, added XEN-tools + XEN-kernel, > enabled bridging, verified with "xm list" that all was up-and-running. > xm/xend is on the way out. xl/libxl (aka libxenlight) is on the way in, and the default Xen toolstack in 13.1 See the following bug for some details, in particular comment#11 https://bugzilla.novell.com/show_bug.cgi?id=857271 I also recently wrote a blog post about libvirt's use of the old and new Xen toolstacks http://jfehlig.wordpress.com/2014/01/05/libvirt-support-for-xens-new-libxenlight-toolstack/ Regards, Jim > But when trying to create a domU (with the yast2-module) no show. > First issue was that an ubuntu-client needed hw-virtualisation (which > was disabled in the bios), but after tweaking that, still no show. > > Next i tried sles11sp3 (thus using paravirt.) but same problem. > Disabled graphic-emulation, still fails. > (sorry for the briefness, no access to those systems here athome) > > Is there anything i should tweak on 13.1? > On 12.1 and 12.2 it worked out-of-the-box... > > Hans > -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] Re: Static IP and internet access in a VM
Sankar P wrote: > 2014/1/7 Bruno Friedmann : > >> On Tuesday 07 January 2014 10.26:34 Sankar P wrote: >> >>> Hi Bruno, >>> >>> Sorry for the late reply as I did not get your mail as I was not >>> subscribed to the >>> list and I forgot to add my email address explicitly in the CC. >>> >>> >>> The qmenu xml definition in a host will look like >>> Also, sorry to say that I was not able to understand what you meant by >>> qmenu xml definition. I am a total newbie when it comes to setting up >>> VMs and have used jsut virtualbox with its helpful gui so far. If it >>> is not too much of an effort, can you explain the steps for >>> using/generating this XML or point me to relevant documentation where >>> I can get static IP for my VM but with internet access ? I searched >>> around but could not find anything newbie friendly. >>> >>> Thanks. >>> >>> Sankar >>> >>> >> Sorry if it was not clear, first I would recommend this bible :-) >> http://activedoc.opensuse.org/book/opensuse-virtualization-with-kvm >> >> You will then learn how to edit manually, or with virsh or virt-manager gui >> the definition of your vm. >> >> Hope this help you, otherwise, just ping back. >> >> > > ah, there seem to be a problem. The doc recommends me to create a new > bridged network card (section 4.1.1.5) . Also my VM currently says: > libvirt connection does not support interface management. What distro? 13.1? Since 12.1, libvirt uses libnetcontrol for interface management on openSUSE. > However, > since I am using network-manager I cannot use YaST to create a new > bridged network card. I want to continue using network manager, as I > want to access my wireless at home etc. > > So what will be the way forward ? To be honest, it feels like such a > basic workflow requirement to require a static IP for a VM, that I am > surprised that it is not straightforward. Any other help or pointers ? > The libvirt wiki has quite a bit of info regarding libvirt's network support http://wiki.libvirt.org/page/Main_Page#Networking Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] Question start order KVM client
Günther J. Niederwimmer wrote: > Hello, > > Is it possible to define the start order for KVM Clients I found nothing in > the > www. > Do you mean the start order when using libvirt-guests to start the VMs? I think there are some hacky tricks, such as naming the domains in a certain order https://www.redhat.com/archives/libvir-list/2011-April/msg00789.html But currently, libvirt doesn't support the notion of VM groups, which is necessary to express ordering constraints. Regards, Jim > Background, for my home server I like a separate Ldap, kerberos dns server, > but when I start this server, the start order is always different. > > Ok I can clone all this on a second server, and I have no Problem with > missing > ldap dns usw. on start. > > But the nicer way is to have all this on one server ;). > > -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] After Update KVM lost his clients
Günther J. Niederwimmer wrote: > Am Mittwoch, 11. Dezember 2013, 09:57:21 schrieb Jim Fehlig: > >> Günther J. Niederwimmer wrote: >> >>> Am Mittwoch, 11. Dezember 2013, 08:41:10 schrieb Jim Fehlig: >>> > > >>>>>>> after the update from 9-10. 12.13 I have lost all my Client in KVM. >>>>>>> >>>>>> Update of what from where? >>>>>> >>>>> only from the Update channel, 13.1 >>>>> >>>>> I mean udev and systemd Update and... >>>>> >> I just applied the latest updates (which included udev, systemd, etc.) >> to one of my 13.1 test machines, rebooted, and see that all of my VMs >> are still there. >> > I have this packets. > > bbs:~ # rpm -qa | grep libvirt > libvirt-python-1.1.2-2.5.1.x86_64 > libvirt-daemon-driver-network-1.1.2-2.5.1.x86_64 > libvirt-daemon-driver-interface-1.1.2-2.5.1.x86_64 > libvirt-client-1.1.2-2.5.1.x86_64 > libvirt-daemon-driver-nwfilter-1.1.2-2.5.1.x86_64 > libvirt-daemon-driver-nodedev-1.1.2-2.5.1.x86_64 > libvirt-daemon-qemu-1.1.2-2.5.1.x86_64 > libvirt-daemon-driver-storage-1.1.2-2.5.1.x86_64 > libvirt-daemon-driver-secret-1.1.2-2.5.1.x86_64 > libvirt-daemon-driver-qemu-1.1.2-2.5.1.x86_64 > libvirt-daemon-1.1.2-2.5.1.x86_64 > Looks sane. I suppose we should continue with this in bugzilla. When starting libvirtd, any output to syslog/journalctl would be helpful. Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] After Update KVM lost his clients
Günther J. Niederwimmer wrote: > Am Mittwoch, 11. Dezember 2013, 08:41:10 schrieb Jim Fehlig: > >> Günther J. Niederwimmer wrote: >> >>> Am Mittwoch, 11. Dezember 2013, 08:09:17 schrieb Jim Fehlig: >>> >>>> Günther J. Niederwimmer wrote: >>>> >>>>> Hello, >>>>> >>>>> after the update from 9-10. 12.13 I have lost all my Client in KVM. >>>>> >>>> Update of what from where? >>>> >>> only from the Update channel, 13.1 >>> >>> I mean udev and systemd Update and... >>> I just applied the latest updates (which included udev, systemd, etc.) to one of my 13.1 test machines, rebooted, and see that all of my VMs are still there. >>> >>>>> the configuration exist but virsh start is no longer working or a >>>>> "virsh list" don't found the clients also in virtual-manager the client >>>>> are gone ? >>>>> >>>> Can virsh event connect to libvirtd? You don't see any of your VMs with >>>> e.g. 'virsh list --all'? Can you get capabilities through virsh, e.g. >>>> 'virsh capabilities'? >>>> >>> with virsh list --all >>> >>> I have also a empty list >>> >> Odd. Do the configuration files for the VMs still exist in >> /etc/libvirt/qemu ? >> > > Yes 5 in /etc/libvirt/qemu and 4 in /etc/kvm/vm ? > So the configuration files are still there and libvirtd should be reading them when started. What libvirt packages do you have installed? Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] After Update KVM lost his clients
Günther J. Niederwimmer wrote: > hello, > > Am Mittwoch, 11. Dezember 2013, 08:09:17 schrieb Jim Fehlig: > >> Günther J. Niederwimmer wrote: >> >>> Hello, >>> >>> after the update from 9-10. 12.13 I have lost all my Client in KVM. >>> >> Update of what from where? >> > > only from the Update channel, 13.1 > > I mean udev and systemd Update and... > > >>> the configuration exist but virsh start is no longer working or a >>> "virsh list" don't found the clients also in virtual-manager the client >>> are gone ? >>> >> Can virsh event connect to libvirtd? You don't see any of your VMs with >> e.g. 'virsh list --all'? Can you get capabilities through virsh, e.g. >> 'virsh capabilities'? >> > > with virsh list --all > > I have also a empty list Odd. Do the configuration files for the VMs still exist in /etc/libvirt/qemu ? > > > with virsh capabilities I have a Long output, but nothing from my Clients? > Ok, so you can connect to libvirtd, but for some strange reason your VM XML configuration files have disappeared? Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] After Update KVM lost his clients
Günther J. Niederwimmer wrote: > Hello, > > after the update from 9-10. 12.13 I have lost all my Client in KVM. > Update of what from where? > the configuration exist but virsh start is no longer working or a "virsh > list" don't found the clients also in virtual-manager the client are gone ? > Can virsh event connect to libvirtd? You don't see any of your VMs with e.g. 'virsh list --all'? Can you get capabilities through virsh, e.g. 'virsh capabilities'? Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] libvirt packaging have now requires for virtualbox ?
Bruno Friedmann wrote: > With the new packaging of the virtual stack, libvirt now push dependencies on > xen, AND virtualbox. > > Look like we are getting more and more stuff auto-bad-gically installed. > I don't need xen, nor I don't want to have virtualbox (which ask > virtualbox-qt, which then ask other things on a server) > > Is it possible (I didn't dig into packaging spec) to make libvirt smarter. > For example if someone want only kvm virtualization he didn't get xen, lxc, > virtualbox. > > In link with yast, there was a time you could choose on virtual system > (kvm,xen). So the yast module has to bee re-aligned with new possible choices. > > Perhaps I've lost track on something. So my assumptions good be wrong, but I > would like to hear someone about the benefit over the inconvenient. > What you are seeing is simply a bug, already found by AJ https://bugzilla.novell.com/show_bug.cgi?id=848918#c5 I spent quite a bit of time cleaning up the upstream spec file (which is the basis for the SUSE spec file) https://www.redhat.com/archives/libvir-list/2013-December/msg00355.html A new maintenance update has been submitted to fix the issue https://build.opensuse.org/request/show/210011 Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] libvirtd fails to start of NetworkManager is enabled
Jim Fehlig wrote: > Olaf Hering wrote: > >> On Sun, Nov 24, Olaf Hering wrote: >> >> >> >>> What is supposed to happen with libvirt if NM is active on a 11.4 >>> system? This used to work at least two weeks ago. >>> >>> >> libvirtd fails also on 13.1. >> > > Ok, I'll take a look. > Should be fixed now. libvirt commit 822fe136 made the interface driver stateful and moved ncf_init to the driver's stateInitialize function. Failure to initialize any of the stateful drivers results in failure to start libvirtd. I changed the stateInitialize function to warn, but still return success when ncf_init fails. Subsequent attempts to use the driver fail as before # virsh iface-list error: Failed to list active interfaces error: this function is not supported by the connection driver: virConnectNumOfInterfaces Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] libvirtd fails to start of NetworkManager is enabled
Olaf Hering wrote: > On Sun, Nov 24, Olaf Hering wrote: > > >> What is supposed to happen with libvirt if NM is active on a 11.4 >> system? This used to work at least two weeks ago. >> > > libvirtd fails also on 13.1. Ok, I'll take a look. > And now that NM starts to handle bridge and > bond, whats libvirt supposed to do with that new functionality? > Nothing, since libvirt has no integration with NM. IMO, libvirt is a peer to NM. So the better question is how are these two supposed to co-exist? We need to have a discussion with the network folks... Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] libvirtd fails to start of NetworkManager is enabled
Olaf Hering wrote: > Current libvirtd from Virtualization/libvirt fails to start because > ncf_init() fails. nc__suse_config_init fails because it expects > NETWORKMANAGER=no, so that "network.service" is returned to > __netcontrol_enabled. > netcontrol was appearently not changed since a long time. So I think for > some reason libvirtd changed. Upstream git did appearently not change, > netcfStateInitialize always returned some error if ncf_init() fails. > Right. And daemonInitialize() in daemon/libvirtd.c ignores the return value anyhow. I'm puzzled why libvirtd won't load when netcfStateInitialize() fails... > What is supposed to happen with libvirt if NM is active on a 11.4 > system? Bad things will happen :-). On a serious note, libvirt and NM are both "network management clients" and may not play well together. I'm not sure if any of the new wicked infrastructure provides better support for multiple network management tools. Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] xen* update DISables `xencommons` and `xendomains` systemd services. should it?
ar16 wrote: > Reminding myself as it happened again this morning ... bump. > Best to open a bug I suppose, so this doesn't get lost. Visual review of the spec file didn't reveal any obvious bug, so this will require more investigation, which is low priority for me ATM - apologies. -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] xen* update DISables `xencommons` and `xendomains` systemd services. should it?
ar16 wrote: > After each xen-libs/tools update, both `xencommons` and `xendomains` > systemd services are DISabled, requiring: > > systemctl enable xencommons.service > ln -s '/usr/lib/systemd/system/xencommons.service' > '/etc/systemd/system/multi-user.target.wants/xencommons.service' > systemctl enable xendomains.service > ln -s '/usr/lib/systemd/system/xendomains.service' > '/etc/systemd/system/multi-user.target.wants/xendomains.service' > > to straighten things out. > > Is that intentional? Are updates typically supposed to modify service > state? > No, I don't think so. Must be a bug in the spec file, which I was unable to quickly spot. My day is about done, so will have to look at this later. -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] is there a clever way to shrink qcow2 images
Rob Verduijn wrote: > Hello all, > > I'm looking for a clever way to shrink qcow2 images. > Thanks for all the info you provided in this thread! I'd like to add that virt-sparsify, virt-resize, and the other virt-* tools in the guestfs-tools package are also useful for manipulating images. Regards, Jim > what I do now is : > > 1 in the vm delete the files I don't need (tempfiles, spoolfiles, > junkfiles, etc, etc) > 2 fill the empty space with zeros > dd if=/dev/zero of=bigfile bs=1024k ; sync ; rm bigfile ; sync > 3 shut down the vm > 4 convert the qcow2 to a fresh new qcow2 > qemu-img convert -c -O qcow2 orig.qcow2 shrunk.qcow2 > mv orig.qcow2 orig.qcow2.bak > mv shrunk.qcow2 orig.qcow2 > 5 test the vm > 6 delete the backup if it works > > Now this is fine if you got plenty space and just want to tidy up a > bit before it turns bad. > > But what if there is no space, and adding extra partitions is not that easy. > (imagine an offsite, far off location that is a real pain to get into > due to serious security hassle) > > Or if you are me and are seriously annoyed by the cumbersome exercise > above and have a gut feeling that there should be a more elegant way > (preferably one without downtime) > > The above exercise requires to be done on all the partitions of the > qcow2 image if there is more than one, adding to my irritation of the > exercise. > > Does anybody have any ideas on a more elegant solution ? > > cheers > Rob > -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] "EVTCHNOP failure" @ OS 12.3 + Xen 4.3 boot for some Guests; launch manually OK
Charles Arnold wrote: On 8/16/2013 at 02:43 PM, in message > <1376685807.15628.10690689.56fd1...@webmail.messagingengine.com>, ar16 > wrote: > >>> I'll experiment without the xendomains.service file and if it seems >>> >> reasonable, I'll remove it. >> >> ok, so we _are_ talking about the xendomains.service now. >> >> back to my point -- imo, fine as long as there's still a >> clear/straightforward way to disable 'just' Guest-startup at boot ... >> > > Now it's my bad. xenstored.service (and xenconsoled.service as well) to be > dropped. > :) Too many xen*.service files, which is more impetus to remove a few of them. -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] "EVTCHNOP failure" @ OS 12.3 + Xen 4.3 boot for some Guests; launch manually OK
ar16 wrote: >> This is probably the case. xenstored.service must be running before >> xencommons.service. Since xencommons will start xenstored, we may not >> need xenstored.service right? >> > > It's useful to be able to disable xen guest launching on boot. > Right, that is the default config. > disabling xendomains.service seems a logical way/place to do that. > Right, not enabled out of the box. > if not, &/or you remove it, what't the 'right' way to do it? > I'm not talking about xendomains.service, but xenstored.service. xencommons starts xenstored, so no need to have systemd do that and hence no need for the service file. -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] "EVTCHNOP failure" @ OS 12.3 + Xen 4.3 boot for some Guests; launch manually OK
Charles Arnold wrote:
>> i'm staring at
>>
>> if ! `${BINDIR}/xenstore-read -s / >/dev/null 2>&1`
>>
>> what exactly is the intent here?
>>
>
> It checks the status of xenstore to see if it is running by reading from the
> socket (-s). On failure, it executes the code to start xenstored. After
> loading xenstored it then hits a while loop checking for up to 30 seconds to
> see if things were properly setup.
> The only failure points are if xenstored is already running,
This is probably the case. xenstored.service must be running before
xencommons.service. Since xencommons will start xenstored, we may not
need xenstored.service right?
--
To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org
To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] "EVTCHNOP failure" @ OS 12.3 + Xen 4.3 boot for some Guests; launch manually OK
ar16 wrote:
> xenstore-ls
> tool = ""
>xenstored = ""
> vm = ""
>www... = ""
> uuid = "www..."
> name = "test2"
> image = ""
> ostype = "linux"
> kernel = "/var/run/xen/bootloader.2.d/boot_kernel.2DQGzh"
> ramdisk = "/var/run/xen/bootloader.2.d/boot_ramdisk.4TDO9l"
> cmdline = "root=/dev/xvdc1 resume=/dev/xvdb1 kbdtype=us
> headless text quiet nofb selinux=0 apparmor-0 edd=off
> splash=silent noshell showopts ide0=noprobe ide1=noprobe
> hda=noprobe root=/dev/xvdc1 textmode=1 xencons=xvc0
> noirqdebu\..."
> start_time = "1376592317.47"
>xxx... = ""
> uuid = "xxx..."
> name = "test2"
> image = ""
> ostype = "linux"
> kernel = "/var/run/xen/bootloader.2.d/boot_kernel.rrFFtJ"
> ramdisk = "/var/run/xen/bootloader.2.d/boot_ramdisk.Sj2nYp"
> cmdline = "root=/dev/xvdc1 resume=/dev/xvdb1 kbdtype=us
> headless text quiet nofb selinux=0 apparmor-0 edd=off
> splash=silent noshell showopts ide0=noprobe ide1=noprobe
> hda=noprobe root=/dev/xvdc1 textmode=1 xencons=xvc0
> noirqdebu\..."
> start_time = "1376592554.77"
>yyy... = ""
> uuid = "yyy..."
> name = "test2"
> image = ""
> ostype = "linux"
> kernel = "/var/run/xen/bootloader.2.d/boot_kernel.8ACINd"
> ramdisk = "/var/run/xen/bootloader.2.d/boot_ramdisk.QKWG2S"
> cmdline = "root=/dev/xvdc1 resume=/dev/xvdb1 kbdtype=us
> headless text quiet nofb selinux=0 apparmor-0 edd=off
> splash=silent noshell showopts ide0=noprobe ide1=noprobe
> hda=noprobe root=/dev/xvdc1 textmode=1 xencons=xvc0
> noirqdebu\..."
> start_time = "1376598384.76"
>zzz... = ""
> uuid = "zzz..."
> name = "test2"
> image = ""
> ostype = "linux"
> kernel = "/var/run/xen/bootloader.2.d/boot_kernel.cc08JL"
> ramdisk = "/var/run/xen/bootloader.2.d/boot_ramdisk.mLPoB3"
> cmdline = "root=/dev/xvdc1 resume=/dev/xvdb1 kbdtype=us
> headless text quiet nofb selinux=0 apparmor-0 edd=off
> splash=silent noshell showopts ide0=noprobe ide1=noprobe
> hda=noprobe root=/dev/xvdc1 textmode=1 xencons=xvc0
> noirqdebu\..."
> start_time = "1376598661.27"
>aaa... = ""
> uuid = "aaa..."
> name = "test2"
> image = ""
> ostype = "linux"
> kernel = "/var/run/xen/bootloader.2.d/boot_kernel.GU6fZ7"
> ramdisk = "/var/run/xen/bootloader.2.d/boot_ramdisk.NyNqNT"
> cmdline = "root=/dev/xvdc1 resume=/dev/xvdb1 kbdtype=us
> headless text quiet nofb selinux=0 apparmor-0 edd=off
> splash=silent noshell showopts ide0=noprobe ide1=noprobe
> hda=noprobe root=/dev/xvdc1 textmode=1 xencons=xvc0
> noirqdebu\..."
> start_time = "1376610136.76"
>bbb... = ""
> uuid = "bbb..."
> name = "test1"
> image = ""
> ostype = "linux"
> kernel = "/var/run/xen/bootloader.1.d/boot_kernel.sYNpOD"
> ramdisk = "/var/run/xen/bootloader.1.d/boot_ramdisk.KjtNA7"
> cmdline = "root=/dev/xvdc1 resume=/dev/xvdb1 kbdtype=us
> headless text quiet nofb selinux=0 apparmor-0 edd=off
> splash=silent noshell showopts ide0=noprobe ide1=noprobe
> hda=noprobe root=/dev/xvdc1 textmode=1 xencons=xvc0
> noirqdebu\..."
> start_time = "1376672796.72"
>ddd... = ""
> uuid = "ddd..."
> name = "test2"
> image = ""
> ostype = "linux"
> kernel = "/var/run/xen/bootloader.2.d/boot_kernel.7xanX5"
> ramdisk = "/var/run/xen/bootloader.2.d/boot_ramdisk.rr6wPE"
> cmdline = "root=/dev/xvdc1 resume=/dev/xvdb1 kbdtype=us
> headless text quiet nofb selinux=0 apparmor-0 edd=off
> splash=silent noshell showopts ide0=noprobe ide1=noprobe
> hda=noprobe root=/dev/xvdc1 textmode=1 xencons=xvc0
> noirqdebu\..."
> start_time = "1376672804.77"
> libxl = ""
>1 = ""
> dm-version = "qemu_xen"
>2 = ""
> dm-version = "qemu_xen"
> local = ""
>domain = ""
> 0 = ""
>
Odd. There should be a name = "Domain-0" node under /local/domain/0, e.g.
# xenstore-ls /local/domain/0
name = "Domain-0"
device-model = ""
...
And that looks to be written by the do_start() function in xencommons.
Maybe you can poke around there and see why
${BINDIR}/xenstore-write "/local/domain/0/name" "Domain-0"
is not being executed.
--
To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org
To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] "EVTCHNOP failure" @ OS 12.3 + Xen 4.3 boot for some Guests; launch manually OK
Charles Arnold wrote: On 8/15/2013 at 05:46 PM, in message > <1376610383.9775.10349293.0d21b...@webmail.messagingengine.com>, ar16 > wrote: > >> doing >> >> edit /etc/tmpfiles.d/xendomains.conf >> D /var/run/xen 0755 root root - >> >> systemd-tmpfiles --create /etc/tmpfiles.d/xendomains.conf >> >> here, does the trick. after reboot, >> >> xl list >> NameID Mem VCPUs >> State Time(s) >> (null) 0 1269 4 >> r- 43.7 >> test11 1024 2 >> -b 7.7 >> test22 1024 4 >> -b 28.6 >> >> and back to full function. >> >> still have that "(null") rather than Dom0 in its Name, but doesn't >> appear (yet) to have any detrimental effect on function >> > > This missing /var/run/xen and evtchn not being loaded seems to indicate that > xencommons is not executed at boot. It modprobes in evtchn and creates > /var/run/xen. > Hmm, we might want to ensure xendomains.service is run After xencommons.service. -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] "EVTCHNOP failure" @ OS 12.3 + Xen 4.3 boot for some Guests; launch manually OK
ar16 wrote: > On Thu, Aug 15, 2013, at 01:47 PM, Jim Fehlig wrote: > >> Agreed, Xen should not depend on the external network being up. But we >> are talking about xendomains, which IMO should depend on external >> network. The domains it starts may need that... >> > > Fair enough. > > At this point, is the add'n of the "network.services" dependency in the > queue to be added to pkgs? Or does a bug need to get filed? > carnold will take care of it, no bug necessary IMO. -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] "EVTCHNOP failure" @ OS 12.3 + Xen 4.3 boot for some Guests; launch manually OK
Jim Fehlig wrote: > ar16 wrote: > >> I'm not sure that Xen *should* depend on external network being up, but >> rather just local network and bridges -- even without attached eth0 >> interfaces. I think ... >> >> > > Agreed, Xen should not depend on the external network being up. But we > are talking about xendomains, which, btw, is not enabled by default and thus no external network dependency when simply booting Xen. Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] "EVTCHNOP failure" @ OS 12.3 + Xen 4.3 boot for some Guests; launch manually OK
ar16 wrote: >> Perhaps the former. Does adding 'network.target' to 'After' in >> xendomains.service help? E.g. >> >> After=xenstored.service xenconsoled.service network.target >> > > For > > ls -al > /etc/systemd/system/multi-user.target.wants/xendomains.service > /usr/lib/systemd/system/xendomains.service > lrwxrwxrwx 1 root root 42 Aug 13 19:34 > /etc/systemd/system/multi-user.target.wants/xendomains.service > -> /usr/lib/systemd/system/xendomains.service > -rw-r--r-- 1 root root 416 Aug 5 03:39 > /usr/lib/systemd/system/xendomains.service > > Adding the systemd dependency on network, > > edit /usr/lib/systemd/system/xendomains.service > - After=xenstored.service xenconsoled.service > + After=xenstored.service xenconsoled.service > network.service > > After reboot > > shutdown -r now > ... > > both Guests, with their complete/original/uncommented .cfg's, launch > again as expected, > > xl list > NameID Mem > VCPUs State Time(s) > (null) 0 1241 > 4 r- 43.3 > test11 1024 > 2 -b 8.4 > test22 1024 > 2 -b 26.8 > > and, checking, seem to be fully functional. > > Looks like adding the 'network.service' dep does the trick. > > Just one question ... > > I'm not sure that Xen *should* depend on external network being up, but > rather just local network and bridges -- even without attached eth0 > interfaces. I think ... > Agreed, Xen should not depend on the external network being up. But we are talking about xendomains, which IMO should depend on external network. The domains it starts may need that... Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] "EVTCHNOP failure" @ OS 12.3 + Xen 4.3 boot for some Guests; launch manually OK
ar16 wrote: > > re: issues with br0 above, checking > > brctl show > bridge name bridge id STP enabled > interfaces > br0 8000.005235d94336 no > eth0 > vifT2.int > > br0 is clearly 'up' and found/used by 'test2' guest. missing systemd > dependency? not waiting long enough @ xendomain start? > Perhaps the former. Does adding 'network.target' to 'After' in xendomains.service help? E.g. After=xenstored.service xenconsoled.service network.target Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] "EVTCHNOP failure" @ OS 12.3 + Xen 4.3 boot for some Guests; launch manually OK
ar16 wrote:
> Ooops, sry, cp'd the wrong result. Same result for BOTH Guests, alone.
>
> My point is that in BOTH cases, the single-guest -- whether it's "test1"
> or "test2" -- launches without error.
>
> If BOTH guests are included to auto-launch, then only the one launches
> successfully.
>
Were you using the xl toolstack before the packages were updated?
Do both guests start if you run the xendomains script after boot? I.e.
disable the script from running at boot, then do '/etc/init.d/xendomains
start' after boot.
Your suspicion of systemd might warrent some investigation
# rpm -q xen --changelog | grep -A 7 systemd
- Added support for systemd with the following service files
xenstored.service
blktapctrl.service
xend.service
xenconsoled.service
xen-watchdog.service
xendomains.service
xencommons.service
Perhaps a bit of tweaking on the xendomains script might give some hints
too, e.g.
--- /etc/init.d/xendomains.orig2013-08-15 11:53:39.487569293 -0600
+++ /etc/init.d/xendomains2013-08-15 11:54:33.440841893 -0600
@@ -34,7 +34,7 @@
$CMD list &> /dev/null
if test $? -ne 0
then
-CMD=${SBINDIR}/xl
+CMD=${SBINDIR}/xl -v
HEADCOMP="Xen saved domain"
fi
@@ -297,7 +297,7 @@
if [ $? -eq 0 ] || is_running $dom; then
echo -n "(skip)"
else
-XMC=`$CMD create --quiet --defconfig $dom`
+XMC=`$CMD create --defconfig $dom`
if [ $? -ne 0 ]; then
echo -e "\nAn error occurred while creating domain
${dom##*/}: $XMC\n"
rc_failed $?
Regards,
Jim
--
To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org
To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] howto change virtual machine creation defaults
Rob Verduijn wrote: > Hello, > > I want to change the defaults when a new vm is created. > ie now the default hd is a raw image that is fully allocated when you > create a vm. > I would like to change it to qcow2 sparse image. > Does anybody know how to do this or where to find documentation for this ? > See /etc/default/vm-install Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] future of XEN (within SuSE)
Hans Witvliet wrote: > Hi all, > > It could be just my perception, but i've got the feeling that more and > more attention is given to KVM, and i wonder if this is intentionally? > > For instance, on http://doc.opensuse.org/ it is about "Virtualization > with KVM" in htmp, pdf and epub, while not a single word about XEN or > LXC... Intentionally? > I don't think it is intentional. AFAIK, the docs come from the SLE docs, and seems no one has added the Xen ones to doc.opensuse.org https://www.suse.com/documentation/sles11/book_xen/?page=/documentation/sles11/book_xen/data/book_xen.html > I got quite some paravirtualized servers, this is seriously worrying me. > And as a long term XEN user, should i start looking for an alternative > distro in case SuSE is going to drop XEN? > There are no plans to drop Xen support. SUSE is one of the largest contributors upstream (2nd in upcoming Xen 4.3 [1] release) and we will continue to provide Xen packages in the future. openSUSE13.1 will have Xen 4.3.x, where the new xl/libxl toolstack will be default and xm/xend is deprecated. Work continues on the Xen front within SUSE... Regards, Jim [1] http://wiki.xen.org/wiki/Xen_4.3_Acknowledgements -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] does anybody know a site with good comparison between all the virtual disk image types ?
Rob Verduijn wrote: > Hello all, > > I'm wondering what the pro's and cons of all the volume format types are ? > what are the (dis)advantages of any of the following ? > raw, bochs, cloop, cow, dmg, iso, qcow, qcow2, qed, vmdk, vpc > I'm not aware of any summary of all the image formats. I think they are all well documented, just no single source of info comparing/contrasting them. > and what would be the pro's and con's to put a certain volume format > on a storage pool type ? > Directory pool, Filesystem pool, Logical volume pools, Disk volume pools, ? > The various libvirt storage pools are described here http://libvirt.org/storage.html > For example : > raw volume format is highly portable, most virtualization solutions > support this type > Yes, that is certainly an advantage of raw. Performance is quite good as well. Downside is that raw is not really an image format like qcow2, qed, and vmdk, so doesn't support cow, snapshot, etc. > nfs is simple to set up as a remote storage pool without huge > investments, but performance may not be optimal. > iscsi is slightly more complex than nfs but also cheap to set up and > when configured properly performs better. > lvm partitions are very usefull > Does an iscsi lun that is a raw partition on the target server > outperform an iscsi lun that is a file on a disk on the target server > ? > I would think so, but haven't done such a performance test. > Does anybody know a site that has a good comparison between the ones > mentioned above. (or some of them) > Unfortunately no, but please do respond if you have found or know of such references. Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] xendomains starts domains although xenstore is not started on opensuse 12.3
Olaf Hering wrote: > On Fri, Mar 15, Dieter Bloms wrote: > > >> Hi, >> >> I've installed opensuse 12.3 as dom0 and after a reboot my first three >> domUs don't start. >> >> In the logfile /var/log/messages I see the following messages, which let >> me assume, that xendomains is started (16:34:27) even xenstore has >> finished starting (16:34:30). >> >> So what can I do to let xendomains start after xenstore has finished >> starting ? >> > > There are two bugs: > One is that systemd appearently does not handle the dependency of > "xenstored xenconsoled" in /etc/init.d/xendomains, which is provided in > /etc/init.d/xencommons. > As a side note, IMO the xendomains script should also depend on $network, since the domains it starts probably need networking. Jim > The second one, likely unrelated to this, is that /etc/init.d/xend also > lacks the RequiredStart: xenstored xenconsoled, and xendomains should > probably also depend on xend. > > If you do these modifications, does xendomains start after xend? > > Olaf > -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] xen vm can't autoboot because of bridge hasn't been set up.
Jun Hu wrote: > 1.though xendomains scripts 'Required-Start' has no '$network" > > but xendomanis depend xend , we can see /etc/init.d/.depend.start : > > network-remotefs: network > ntp: network-remotefs syslog > xend:nt p > xendomains: xend > > so , apparently , there is no problem about dependency relationship. > Right. > 2. after I add '$network" in 'Required-Start' line of xendomains > scripts, > run insserv xendomains > > reboot , but found the issue exists still. > Not surprising, given your point above. > 3. on opensuse, it use systemd instead of sysvinit , I think cause > may be there. > Agreed. IMO, you should file a bug against systemd. Proper systemd service files for the xen package would be nice too. Regards, Jim > > > On Thu, 2013-01-03 at 18:09 -0700, Jim Fehlig wrote: > >> Jun Hu wrote: >> >>> Hi Guys: >>> >>> I found an issue that xen vm can't autoboot because of bridge hasn't >>> set up, i can manually boot the vm . >>> >>> maybe a bug? >>> >> Seems the xendomains script should require the network to be active. >> Does adding '$network' to 'Required-Start' and 'Required-Stop' help? >> >> Regards, >> Jim >> >> >>> >>> >>> junwork:/etc/libvirt/qemu # chkconfig -A --list | grep xen >>> >>> Note: This output shows SysV services only and does not include native >>> systemd services. SysV configuration data might be overridden by native >>> systemd configuration. >>> >>> xen-watchdog 0:off 1:off 2:off 3:off 4:off 5:off >>> 6:off >>> xencommons0:off 1:off 2:on 3:on 4:off 5:on >>> 6:off >>> xend 0:off 1:off 2:off 3:on 4:off 5:on >>> 6:off >>> xendomains0:off 1:off 2:off 3:on 4:off 5:on >>> 6:off >>> >>> >>> junwork:~ # rcxend status >>> redirecting to systemctl >>> xend.service - LSB: Starts and stops the Xen management daemon >>> Loaded: loaded (/etc/init.d/xend) >>> Active: active (running) since Fri, 04 Jan 2013 06:41:11 +0800; 1h >>> 14min ago >>> Process: 659 ExecStart=/etc/init.d/xend start (code=exited, >>> status=0/SUCCESS) >>> CGroup: name=systemd:/system/xend.service >>> ├ 737 /usr/bin/python /usr/sbin/xend start >>> ├ 738 blktapctrl >>> └ 740 /usr/bin/python /usr/sbin/xend start >>> >>> Jan 04 06:41:11 junwork xend[659]: Starting xend ..done >>> >>> >>> junwork:/etc/libvirt/qemu # rcxendomains status >>> redirecting to systemctl >>> xendomains.service - LSB: Starts and stops Xen VMs >>> Loaded: loaded (/etc/init.d/xendomains) >>> Active: failed (Result: exit-code) since Fri, 04 Jan 2013 06:41:31 >>> +0800; 1h 20min ago >>> Process: 1323 ExecStart=/etc/init.d/xendomains start (code=exited, >>> status=1/FAILURE) >>> CGroup: name=systemd:/system/xendomains.service >>> >>> Jan 04 06:41:11 junwork xendomains[1323]: Restoring saved Xen domains >>> Jan 04 06:41:29 junwork xendomains[1323]: windowsxp: ..failed >>> Jan 04 06:41:29 junwork xendomains[1323]: Starting auto Xen domains >>> Jan 04 06:41:31 junwork xendomains[1323]: config_winxp: Error: Device 0 >>> (vif) could not be connected. Could not find ...ce br0 >>> Jan 04 06:41:31 junwork xendomains[1323]: ..failed >>> >>> junwork:/etc/libvirt/qemu # ifconfig >>> br0 Link encap:Ethernet HWaddr 00:1F:16:25:AA:66 >>> inet addr:192.168.5.20 Bcast:192.168.5.255 >>> Mask:255.255.255.0 >>> UP BROADCAST MULTICAST MTU:1500 Metric:1 >>> RX packets:0 errors:0 dropped:0 overruns:0 frame:0 >>> TX packets:0 errors:0 dropped:0 overruns:0 carrier:0 >>> collisions:0 txqueuelen:0 >>> RX bytes:0 (0.0 b) TX bytes:0 (0.0 b) >>> >>> eth0 Link encap:Ethernet HWaddr 00:1F:16:25:AA:66 >>> UP BROADCAST MULTICAST MTU:1500 Metric:1 >>> RX packets:0 errors:0 dropped:0 overruns:0 frame:0 >>> TX packets:0 errors:0 dropped:0 overruns:0 carrier:0 >>> collisions:0 txqueuelen:1000 >>> RX bytes:0 (0.0 b) TX bytes:0 (0.0 b) >>>
Re: [opensuse-virtual] xen vm can't autoboot because of bridge hasn't been set up.
Jun Hu wrote: > Hi Guys: > > I found an issue that xen vm can't autoboot because of bridge hasn't > set up, i can manually boot the vm . > > maybe a bug? Seems the xendomains script should require the network to be active. Does adding '$network' to 'Required-Start' and 'Required-Stop' help? Regards, Jim > > > junwork:/etc/libvirt/qemu # chkconfig -A --list | grep xen > > Note: This output shows SysV services only and does not include native > systemd services. SysV configuration data might be overridden by native > systemd configuration. > > xen-watchdog 0:off 1:off 2:off 3:off 4:off 5:off > 6:off > xencommons0:off 1:off 2:on 3:on 4:off 5:on > 6:off > xend 0:off 1:off 2:off 3:on 4:off 5:on > 6:off > xendomains0:off 1:off 2:off 3:on 4:off 5:on > 6:off > > > junwork:~ # rcxend status > redirecting to systemctl > xend.service - LSB: Starts and stops the Xen management daemon > Loaded: loaded (/etc/init.d/xend) > Active: active (running) since Fri, 04 Jan 2013 06:41:11 +0800; 1h > 14min ago >Process: 659 ExecStart=/etc/init.d/xend start (code=exited, > status=0/SUCCESS) > CGroup: name=systemd:/system/xend.service > ├ 737 /usr/bin/python /usr/sbin/xend start > ├ 738 blktapctrl > └ 740 /usr/bin/python /usr/sbin/xend start > > Jan 04 06:41:11 junwork xend[659]: Starting xend ..done > > > junwork:/etc/libvirt/qemu # rcxendomains status > redirecting to systemctl > xendomains.service - LSB: Starts and stops Xen VMs > Loaded: loaded (/etc/init.d/xendomains) > Active: failed (Result: exit-code) since Fri, 04 Jan 2013 06:41:31 > +0800; 1h 20min ago >Process: 1323 ExecStart=/etc/init.d/xendomains start (code=exited, > status=1/FAILURE) > CGroup: name=systemd:/system/xendomains.service > > Jan 04 06:41:11 junwork xendomains[1323]: Restoring saved Xen domains > Jan 04 06:41:29 junwork xendomains[1323]: windowsxp: ..failed > Jan 04 06:41:29 junwork xendomains[1323]: Starting auto Xen domains > Jan 04 06:41:31 junwork xendomains[1323]: config_winxp: Error: Device 0 > (vif) could not be connected. Could not find ...ce br0 > Jan 04 06:41:31 junwork xendomains[1323]: ..failed > > junwork:/etc/libvirt/qemu # ifconfig > br0 Link encap:Ethernet HWaddr 00:1F:16:25:AA:66 > inet addr:192.168.5.20 Bcast:192.168.5.255 > Mask:255.255.255.0 > UP BROADCAST MULTICAST MTU:1500 Metric:1 > RX packets:0 errors:0 dropped:0 overruns:0 frame:0 > TX packets:0 errors:0 dropped:0 overruns:0 carrier:0 > collisions:0 txqueuelen:0 > RX bytes:0 (0.0 b) TX bytes:0 (0.0 b) > > eth0 Link encap:Ethernet HWaddr 00:1F:16:25:AA:66 > UP BROADCAST MULTICAST MTU:1500 Metric:1 > RX packets:0 errors:0 dropped:0 overruns:0 frame:0 > TX packets:0 errors:0 dropped:0 overruns:0 carrier:0 > collisions:0 txqueuelen:1000 > RX bytes:0 (0.0 b) TX bytes:0 (0.0 b) > Interrupt:20 Memory:f260-f262 > > loLink encap:Local Loopback > inet addr:127.0.0.1 Mask:255.0.0.0 > inet6 addr: ::1/128 Scope:Host > UP LOOPBACK RUNNING MTU:16436 Metric:1 > RX packets:10 errors:0 dropped:0 overruns:0 frame:0 > TX packets:10 errors:0 dropped:0 overruns:0 carrier:0 > collisions:0 txqueuelen:0 > RX bytes:1212 (1.1 Kb) TX bytes:1212 (1.1 Kb) > > wlan0 Link encap:Ethernet HWaddr 00:1E:65:6E:90:AE > inet addr:192.168.1.2 Bcast:192.168.1.255 Mask:255.255.255.0 > inet6 addr: fe80::21e:65ff:fe6e:90ae/64 Scope:Link > UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 > RX packets:6695 errors:0 dropped:0 overruns:0 frame:0 > TX packets:5897 errors:0 dropped:0 overruns:0 carrier:0 > collisions:0 txqueuelen:1000 > RX bytes:3861362 (3.6 Mb) TX bytes:1266239 (1.2 Mb) > > > > junwork:/etc/xen/auto # cat config_winxp > name="windowsxp" > description="ip_192.168.10.55" > uuid="d090b5fe-c5e2-e8fd-6720-a49d5debb94c" > memory=512 > maxmem=512 > vcpus=2 > on_poweroff="destroy" > on_reboot="restart" > on_crash="destroy" > localtime=1 > keymap="en-us" > > builder="hvm" > device_model="/usr/lib/xen/bin/qemu-dm" > kernel="/usr/lib/xen/boot/hvmloader" > boot="c" > disk=[ > 'file:/xendisk/xen_vms/winxp/winxp.img,hda,w', > > ] > vif=[ 'mac=00:16:3e:63:af:b9,bridge=br0,model=rtl8139', > # 'mac=00:16:3e:13:63:2b,bridge=br1,model=rtl8139', > ] > ##ip=192.168.5.221 > > stdvga=0 > vnc=1 > vncunused=1 > extid=0 > usb=1 > acpi=1 > pae=1 > > usbdevice='tablet' > > serial="pty" > > > > -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] Re: KVM weirdness on laptop - Need opnions
Tony Su wrote: > Re -reading my post, there are some small language issues which might > make the post unclear, hoping with the following alterations the post > might be more clear... > > Hello, > > Am hoping to get experienced eyes on this to evaluate how much of what > I'm seeing is expected and what is not. > > System: openSUSE 12.2 > Desktop: KDE 4.8 > Virtualization: QEMU-KVM > Also using graphical virt-manager > > Hardware: Laptop with occasionally available WiFi > > - I have found that anything configured for bridged networking still > requires a wired connection, usually binding to eth0. > > - I have found that when any real connection to a Network exists > (wireless or wired), if a Host Only network does not exist, then the > Guest will be automatically configured to use User Mode Networking > . If any network available, then > User Mode Networking no longer is an option initially or later and the > virtual NIC is configurable for any available virtual network. > Maybe you should try the "default" virtual network provided by libvirt. For more details, see http://wiki.libvirt.org/page/VirtualNetworking The NAT mode works well for me in similar scenarios. Regards, Jim > - If a Guest is initially created with User Mode Networking and then > the Host is disconnected from any and all real physical networks, User > Mode > Networking will fail with an error "service unavailable, port 5900" > which likely means that a VNC connection cannot be made. > > - Curiously, working from the previous unbootable Guest(no service > unavailable, port 5900 error), if a Host Only network is created and a > second virtual network for the new Host Only network is added to the > Guest described previously, the Guest VM will > boot up and can even be > viewed using virt-viewer. But, on closer inspection running ifconfig > in the Guest reveals only one working NIC configured with an IP > address for the User Mode Network while running ifconfig on the Host > reveals that only the Host Only network is running and available. So, > of course, Host and Guest cannot communicate using TCP/IP with each > other except using virt-viewer (should not work?). > > - Have not tried, but am considering the idea of configuring the Host > Only network to share the same networkID as the User Mode Network > (10.0.2.0/24) but make sure the DHCP ranges are different. > Theoretically I think this workaround might work but shouldn't be > required. > > TIA, > Tony > > On Mon, Nov 26, 2012 at 8:33 AM, Tony Su wrote: > >> Hello, >> >> Am hoping to get experienced eyes on this to evaluate how much of what I'm >> seeing is expected and what is not. >> >> System: openSUSE 12.2 >> Desktop: KDE 4.8 >> Virtualization: QEMU-KVM >> Also using graphical virt-manager >> >> Hardware: Laptop with occasionally available WiFi >> >> - I have found that anything configured for bridged networking still >> requires a wired connection, usually binding to eth0. >> >> - I have found that when any real connection to a Network exists (wireless >> or wired), if a Host Only network does not exist, then the Guest will be >> automatically configured to use User Mode Networking > created>. >> >> - When a physical connection does not exist for a Guest previously created >> using a NIC configured for User Mode Networking, User Mode Networking will >> fail with an error "service unavailable, port 5900" which likely means that >> a VNC connection cannot be made. >> >> - Curiously, even with no physical network connection, if a Host Only >> network is created and a second virtual network for the new Host Only >> network is added to the Guest described previously, the Guest VM will boot >> up (no service unavailable, port 5900 error) and can even be viewed using >> virt-viewer. But, on closer inspection running ifconfig in the Guest >> reveals only one working NIC configured with an IP address for the User >> Mode Network while running ifconfig on the Host reveals that only the Host >> Only network is running and available. So, of course, Host and Guest cannot >> -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] outdated docs, kvm/libvirtd auth changes in 12.2
Olaf Hering wrote: > While reading through the KVM documentation I noticed that the current > version of openSuSE does not have the referenced polkit-auth anymore: > > http://doc.opensuse.org/documentation/html/openSUSE/opensuse-kvm/cha.libvirt.connect.html > Yeah, that applies to SLES11, but not for openSUSE since 11.2 or so. I created a comment about this outdated info as per the feedback instructions in "About This Manual". > Has polkit-auth a replacement in 12.2? > polkit-auth is part of the PolicyKit package, which was replaced with polkit in openSUSE long ago. Better documentation for using polkit is here http://wiki.libvirt.org/page/SSHPolicyKitSetup > For my own testing with libguestfs all it takes is to add the user to the > group 'kvm' for the attach-method 'libvirt', so I dont really need > polkit-auth. > Your normal user can do stuff like 'virsh -c qemu+ssh:///system' without configuring polkit as per the libvirt wiki? On my 12.2 machine, I must add the following to authenticate to the system URI (even when user jfehlig is a member of libvirt, kvm, and qemu groups) # cat /etc/polkit-1/localauthority/50-local.d/50-libvirt-remote-access.pkla [Remote libvirt SSH access] Identity=unix-user:jfehlig Action=org.libvirt.unix.manage ResultAny=yes ResultInactive=yes ResultActive=yes -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] Which Virtual machine for PCI-Express graphic card passthrough ?
Hans Witvliet wrote: > On Wed, 2012-10-03 at 13:45 +0200, Dsant wrote: > >> In short : Which Virtual machine does PCI-Express graphic card passthrough ? >> Xen, KVM, VIrtualBox, lxc containers ? >> >> Long : I would like to run proprietary Linux games in a guest (host and >> guest >> Linux). Which virtual machine really works for 3D ? The graphic card can be >> cut off from the host, it's not a problem. >> >> Dsant >> > > Never done it myself, > > But last couple of years i've seen more than a couple of messages on the > xen-list regarding doing pci-passthrough of one (or multiple) vga-cards > towards domU's. (both para- and full-virtualisation) > Although I haven't personally tried, there has been some success with GPU passthrough on Xen. IIRC, success is more likely with some cards than others. E.g. the Intel cards work pretty well. > Can't tell about kvm or virtualbox, but for LXC it seems unlikely... > I think the same can be said for KVM, but also don't know about VirtualBox. Definitely not possible with LXC. Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] virt-manager unable to connect to libvirt after update
Jim Fehlig wrote: > Ralf Müller wrote: > >> Am 29.08.2012 um 17:34 schrieb Jim Fehlig: >> >> >> >>> For some reason, libvirt is insisting on using the qemu driver. What if >>> you explicitly specify xen? E.g. 'virsh -c xen:/// version'? >>> >>> >> # virsh -c xen:/// version >> WARNING: gnome-keyring:: no socket to connect to >> error: internal error libxenlight state driver is not active >> error: failed to connect to the hypervisor >> >> >> >>> I'll try to reproduce this too. >>> >>> >> Tell me if you need information to do that. >> >> > > I'm able to reproduce and looking into it now. > This turns out to be quite a mess with all the recent upstream changes wrt loadable driver modules. I'll be working with upstream to get this resolved. Stay tuned... > Would you have time to open a bug for tracking? > Please open a bug. Thanks for reporting! Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] virt-manager unable to connect to libvirt after update
Ralf Müller wrote: > Am 29.08.2012 um 17:34 schrieb Jim Fehlig: > > >> For some reason, libvirt is insisting on using the qemu driver. What if >> you explicitly specify xen? E.g. 'virsh -c xen:/// version'? >> > > # virsh -c xen:/// version > WARNING: gnome-keyring:: no socket to connect to > error: internal error libxenlight state driver is not active > error: failed to connect to the hypervisor > > >> I'll try to reproduce this too. >> > > Tell me if you need information to do that. > I'm able to reproduce and looking into it now. Would you have time to open a bug for tracking? Thanks, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] virt-manager unable to connect to libvirt after update
Ralf Müller wrote: > Am 29.08.2012 um 16:36 schrieb Jim Fehlig: > > >> Ralf Müller wrote: >> >>> I'm running a Xen-Server using the OpenSuSE Tumbleweed repository. After I >>> updated this system the day before (zypper dup --from Tumbleweed; zypper >>> up; reboot) virt-manager is not able to connect to libvirt anymore. All >>> domU's are up and running, "xm list" shows them correctly, but virt-manager >>> reports the following after start: >>> >>> Unable to connect to libvirt. >>> >>> internal error libxenlight state driver is not active >>> >>> >> That is odd, since it appears you are using xend. When xend is active, >> the legacy libvirt xen driver should be used, not the libxenlight driver. >> > > Can I configure that somewhere? > > >> xend is running correct? >> > > seems to be ok: > # rcxend status > redirecting to systemctl > xend.service - LSB: Starts and stops the Xen management daemon > Loaded: loaded (/etc/init.d/xend) > Active: active (running) since Tue, 28 Aug 2012 08:48:30 +0200; 1 day > and 8h ago >Process: 1766 ExecStart=/etc/init.d/xend start (code=exited, > status=0/SUCCESS) > CGroup: name=systemd:/system/xend.service > ├ 1998 blktapctrl > ├ 2000 /usr/bin/python /usr/sbin/xend start > ├ 2008 /usr/bin/python /usr/sbin/xend start > ├ 4026 /usr/lib/xen/bin/qemu-dm -d 2 -serial pty > -domain-na... > ├ 5695 /usr/lib/xen/bin/qemu-dm -d 3 -serial pty > -domain-na... > ├ 6051 /usr/lib/xen/bin/qemu-dm -d 4 -domain-name gate > -vid... > ├ 8370 /usr/lib/xen/bin/qemu-dm -d 6 -serial pty > -domain-na... > ├ 9037 /usr/lib/xen/bin/qemu-dm -d 7 -serial pty > -domain-na... > ├ 9758 /usr/lib/xen/bin/qemu-dm -d 8 -serial pty > -domain-na... > └ 15750 /usr/lib/xen/bin/qemu-dm -d 9 -serial pty > -domain-na... > > >> What is the output of 'virsh version'? >> > > Thats fascinating: > # virsh version > WARNING: gnome-keyring:: no socket to connect to > Compiled against library: libvir 0.9.13 > Using library: libvir 0.9.13 > Using API: QEMU 0.9.13 > error: failed to get the hypervisor version > error: internal error Cannot find suitable emulator for x86_64 > > What does this say to me? > For some reason, libvirt is insisting on using the qemu driver. What if you explicitly specify xen? E.g. 'virsh -c xen:/// version'? I'll try to reproduce this too. Jim > >> Also, do you see the same error >> when using virsh? E.g. 'virsh list --all'? >> > > # virsh list --all > WARNING: gnome-keyring:: no socket to connect to > IdName State > > > > >> You can try restarting libvirtd, and peek at >> /var/log/libvirt/libvirtd.log to see the results of probing which xen >> driver to use. >> > > # rclibvirtd restart; tail /var/log/messages > redirecting to systemctl > ... > Aug 29 16:58:09 tower-core libvirtd[19796]: 2012-08-29 14:58:09.976+: > 19807: info : libvirt version: 0.9.13 > Aug 29 16:58:09 tower-core libvirtd[19796]: 2012-08-29 14:58:09.976+: > 19807: warning : virSecurityManagerNew:136 : Configured security driver > "none" disables default policy to create confined guests > Aug 29 16:59:00 tower-core libvirtd[19882]: WARNING: gnome-keyring:: no > socket to connect to > > Best Regards > Ralf > > > > -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] virt-manager unable to connect to libvirt after update
Ralf Müller wrote:
> This is a resend from opens...@opensuse.org and xen-us...@lists.xen.org
>
> Hi there.
>
> I'm running a Xen-Server using the OpenSuSE Tumbleweed repository. After I
> updated this system the day before (zypper dup --from Tumbleweed; zypper up;
> reboot) virt-manager is not able to connect to libvirt anymore. All domU's
> are up and running, "xm list" shows them correctly, but virt-manager reports
> the following after start:
>
> Unable to connect to libvirt.
>
> internal error libxenlight state driver is not active
>
That is odd, since it appears you are using xend. When xend is active,
the legacy libvirt xen driver should be used, not the libxenlight driver.
xend is running correct?
What is the output of 'virsh version'? Also, do you see the same error
when using virsh? E.g. 'virsh list --all'?
You can try restarting libvirtd, and peek at
/var/log/libvirt/libvirtd.log to see the results of probing which xen
driver to use.
Jim
> Verify that:
> - A Xen host kernel was booted
> - The Xen service has been started
>
> Details
> Unable to connect to libvirt.
>
> internal error libxenlight state driver is not active
>
> Verify that:
> - A Xen host kernel was booted
> - The Xen service has been started
>
> Libvirt URI is: xen:///
>
> Traceback (most recent call last):
> File "/usr/share/virt-manager/virtManager/connection.py", line 1027, in
> _open_thread
> self.vmm = self._try_open()
> File "/usr/share/virt-manager/virtManager/connection.py", line 1009, in
> _try_open
> flags)
> File "/usr/lib64/python2.7/site-packages/libvirt.py", line 102, in openAuth
> if ret is None:raise libvirtError('virConnectOpenAuth() failed')
> libvirtError: internal error libxenlight state driver is not active
>
> rclibvirtd status reports:
> libvirtd.service - Virtualization daemon
> Loaded: loaded (/lib/systemd/system/libvirtd.service; enabled)
> Active: active (running) since Tue, 28 Aug 2012 09:19:18 +0200;
> 33min ago
> Main PID: 11360 (libvirtd)
> CGroup: name=systemd:/system/libvirtd.service
> └ 11360 /usr/sbin/libvirtd --listen
>
> # uname -a
> Linux tower-core 3.5.2-39-xen #1 SMP Wed Aug 15 21:49:59 UTC 2012 (4904750)
> x86_64 x86_64 x86_64 GNU/Linux
>
> # rpm -q xen
> xen-4.1.2_17-1.10.1.x86_64
>
> Any idea what goes wrong?
>
> Best regards
> Ralf
>
>
--
To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org
To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] RE: PV driver problem about Opensuse 12.1 xen
Hao, Xudong wrote: > Hi, all > > Does anyone work on it and know the solution of this issue? > This question was answered on an internal mailing list, but opensuse-virtual was not included. Here is Olaf's answer: 12.1 has no PV drivers in the inst-sys, in addition the ata_piix driver is compiled into the kernel, and the installer uses kexec to boot from first stage of installation into second stage of installation. The sum of all that leads to the issues described below, in a HVM guest. 12.2 has these issues fixed, as of current Factory. Olaf >> -Original Message- >> From: Hao, Xudong >> Sent: Monday, June 18, 2012 4:23 PM >> To: opensuse-virtual@opensuse.org >> Subject: PV driver problem about Opensuse 12.1 xen >> >> Hi, >> >> We installed Opensuse 12.1 Xen and create Opensuse 12.1 as guest, but guest >> booting has I/O error and system can't boot correctly. I checked the code and >> find qemu unplug IDE disk and try to use PV driver disk. We used default >> configuration for opensuse installation. >> >> Now we disable pv driver in qemu, but it's not the default configuration for >> user. >> Does Opensuse 12.1 use PV driver by default? Does user need some special >> configuration for it, or any solution? >> >> Copy relative code: >> >> =PV driver: >> static int __devinit platform_pci_init(struct pci_dev *pdev, >> ... >> /* >> * Disconnect the emulated devices. >> */ >> outl(1, (ioaddr + 4)); >> ... >> >> >> ==Qemu =: >> static void platform_ioport_write(void *opaque, uint32_t addr, uint32_t val) >> ... >> case 4: >> fprintf(logfile, "Disconnect IDE hard disk...\n"); >> ide_unplug_harddisks(); >> fprintf(logfile, "Disconnect SCSI hard disk...\n"); >> pci_unplug_scsi(); >> ... >> >> >> Best Regards, >> Xudong Hao >> >> -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] python-SpiceClientGtk
Dominique Leuenberger a.k.a DimStar wrote: > > Quoting Jim Fehlig : > >> Rob Verduijn wrote: >>> Hello, >>> I see that python-SpiceClientGtk is published in >>> http://download.opensuse.org/repositories/Virtualization/openSUSE_Factory/ >>> >>> but not in >>> http://download.opensuse.org/repositories/Virtualization/openSUSE_12.1/ >>> >>> Is there a special reason for this ? >>> >> >> It is only enabled for Factory, but a quick look at the spec file shows >> no dependencies that restrict it to Factory. >> >> I would be happy to enable it for 12.1, but would first like to check >> with dimstar, the maintainer of spice-gtk. >> >> Regards, >> Jim > > Jim, > > No objections from my end. I enabled build for 12.1 but noticed several unresolved dependencies libcacard >= 0.1.2 libusb-1.0 >= 1.0.9 libusbredirhost >= 0.3.3 Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] python-SpiceClientGtk
Rob Verduijn wrote: > Hello, > I see that python-SpiceClientGtk is published in > http://download.opensuse.org/repositories/Virtualization/openSUSE_Factory/ > but not in > http://download.opensuse.org/repositories/Virtualization/openSUSE_12.1/ > > Is there a special reason for this ? > It is only enabled for Factory, but a quick look at the spec file shows no dependencies that restrict it to Factory. I would be happy to enable it for 12.1, but would first like to check with dimstar, the maintainer of spice-gtk. Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] what repo to use ?
Rob Verduijn wrote: > Hello, > > Does anybody know what the difference is between these three repos ? > > http://download.opensuse.org/repositories/Virtualization:/openSUSE12.1/openSUSE_12.1_Update_standard > http://download.opensuse.org/repositories/Virtualization:/openSUSE12.1/openSUSE_12.1/ > These two repos come from the OBS Virtualization:openSUSE12.1 project, which is where 12.1 packages are maintained. The first one contains packages built against the latest 12.1 updates. The latter contains packages built against 12.1 GA. Honestly, I'm not sure we even need the latter. > http://download.opensuse.org/repositories/Virtualization/openSUSE_12.1/ > This it the 12.1 repo for the Virtualization project, which is the devel project for Factory. > I find them a bit confusing since I find an older version of libvirt > in the update standard repo than in the opensuse 12.1 repo. > But it is the same code. I suppose the package in the update standard repo has been built fewer times than the one in the GA repo. Hmm, come to think of it, the update repo was added later, so the package hasn't been built against it as often. > Also the difference between the Virtualization: and Virtualization is > a bit fuzy to me. > As mentioned, Virtualization is the devel project for Factory. All the latest and greatest virtualization-related packages are developed and tested in this project before submitting to Factory. Once a product has been released, we create a product subproject under Virtualization, e.g. Virtualization:openSUSE11.4, Virtualization:openSUSE12.1, etc. Maintenance of the virtualization-related packages that shipped with the product are done in these subprojects. > Which of these repos is the one that should be used ? > Depends :-). If you want to use/test the latest virtualization packages, then you would use the ones from a Virtualization project repo. If you want to use/test the latest updated packages for a released product, then use the Virtualization: subproject. Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] Xen kernel not booting openSUSE 12.1
Jan Beulich wrote: On 20.03.12 at 22:06, Sanidhya Kashyap wrote: >> Hi, >> >> I installed the xen kernel for the virtualization support. But the xen >> kernel does not boots up. It hangs for sometime at the time of booting >> and then the system restarts. You can use the 'noreboot' option on xen cmd line to prevent restarting. It is useful for debugging these sorts of problems. >> >> >> the configuration of menu.lst is as follows >> >> >> title Xen -- openSUSE 12.1 - 3.1.9-1.4 >> root (hd0,5) >> kernel /boot/xen.gz vga=mode-normal >> module /boot/vmlinuz-3.1.9-1.4-xen >> root=/dev/disk/by-id/ata-WDC_WD5000BEVT-00A0RT0_WD-WXD1AB0L7670-part6 >> ip=dhcp -- >> resume=/dev/disk/by-id/ata-WDC_WD5000BEVT-00A0RT0_WD-WXD1AB0L7670-part9 >> splash=silent quiet showopts vga=normal >> module /boot/initrd-3.1.9-1.4-xen >> >> I also tried using nomodeset option. But, that also did not work. >> >> So, can anyone tell me what is wrong? Any help would be appreciated! >> > > Without you providing kernel and hypervisor logs or any other > technical information that might help identify _what_ is going > wrong, we're unlikely to be able to help you. > Sanidhya, fyi, one method of getting some helpful info http://old-en.opensuse.org/How_to_Capture_Xen_Hypervisor_and_Kernel_Messages_using_a_Serial_Cable Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] openSUSE 12.1 libvirtd / kvm kernel call trace
Bruno Friedmann wrote: > Hi all, > > I'm now using on a AMD X2 6400 processor openSUSE 12.1 64bits with kvm and > libvirtd > Running a openSUSE factory vm on it drive to getting a lot of messages like > those > > [ 760.739641] kvm: 6261: cpu0 unhandled rdmsr: 0xc0010112 > [ 760.739717] kvm: 6261: cpu0 unhandled rdmsr: 0xc0010048 > [ 760.847212] kvm: 6261: cpu0 unhandled rdmsr: 0xc0010001 > [ 760.875378] kvm: 6261: cpu1 unhandled rdmsr: 0xc0010048 > [ 1336.060011] BUG: soft lockup - CPU#0 stuck for 23s! [qemu-kvm:6261] > [...] > Should I report it as bug ? > Yeah, I think filing a bug would be best. Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] Naming the virtual interface?
Per Jessen wrote: > I've been googling a lot looking for how to use the vifname parameter to > give me pre-defined interface names - there seems to be many hits > discussing whether or not it is supported, but it is certainly tested > in /etc/xen/scripts/vif-common.sh. > > I've got a vif entry with "vifname=guest112.0", but this appears to be > just ignored. > I assume you are using the xl/libxl toolstack? I see the same with it, but noticed vifname is honoured when using xm/xend toolstack. Can you enter a bug please? Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] After switching from "xm" to "xl" toolstack, can't get Guest networking to work.
Per Jessen wrote: > erin.ba...@inoutbox.com wrote: > > >> Hi All. >> >> This problem's got a little - not a lot! - of traction over at the >> xen-devel list. >> >> http://lists.xen.org/archives/html/xen-devel/2012-01/msg01754.html. >> >> Rather than leave this open & uncommented here, I'm going to >> concentrate on that thread instead for now. >> > > Just fyi - I've got a working setup with xl. I'm using openSUSE 12.1 > with xenutils/-tools installed from source though. There is a patch in openSUSE12.1 xen package causing this issue. More info on the xen-devel thread mentioned above. Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] Trying to launching any X11 applications from a Xen Guest on my local desktop doesn't open a window, and fails with an error.
rogerhalb...@operamail.com wrote: > On Tue, Jan 10, 2012, at 11:59 AM, Jim Fehlig wrote: > >> What is the ssh client used here? >> > > which ssh > /usr/bin/ssh > > rpm -q --whatprovides `which ssh` > openssh-5.9p1-20.2.x86_64 > My client is 5.8p2-3.1.3 > I regularly, successfully use this client to x-forward to and from > non-xen boxes, with full X environments, and real video hardware. > Hmm, I can't reproduce. This sure smells like a configuration issue, but unfortunately I'm not familiar with debugging ssh. Maybe one of the user lists (opens...@opensuse.org) would be more helpful. Meanwhile, some debug output might point in the right direction, e.g. 'ssh -X -vvv ...'. Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] Trying to launching any X11 applications from a Xen Guest on my local desktop doesn't open a window, and fails with an error.
rogerhalb...@operamail.com wrote: >> That's exactly what I have been doing. The error I get is in one of >> those "ssh -X" forwarded sessions. >> > > Here's the transcript of my attempt. > > [rh laptop: /home/rh]$ ssh -X -l rh domU_s2.internal.loc > X11 forwarding request failed on channel 0 > What is the ssh client used here? -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] Trying to launching any X11 applications from a Xen Guest on my local desktop doesn't open a window, and fails with an error.
rogerhalb...@operamail.com wrote: > Good morning, > > I'm consolidating a number of separate, physical servers onto a single > Xen platform. > > I have installed OpenSUSE 12.1 as the Xen Host, and, I can install & > launch OpenSUSE 12.1 Xen Guests. > > All of the Guests are accessible using SSH, and are generally > functioning as servers as they're intended to. > > So far, so good. > > For one of the Guests, I need access to a root X11 application on it. > > From what I've managed to read, I understand I can do that by "enabling > X forwarding" on the Guest, opening an SSH session to the Guest, and > executing the target application (e.g. 'xterm'), and if all goes well, > the application's window should appear on (or is it "in"?) my local > desktop. > > Problem is, when I try that, I get the following error in the SSH > console window: > > "xterm > Warning: This program is an suid-root program or is being run by the > root user. > The full text of the error or warning message cannot be safely formatted > in this environment. You may get a more descriptive message by running > the > program as a non-root user or by removing the suid bit on the > executable. > xterm Xt error: Can't open display: %s" > Seems you don't have X forwarding enabled. To enable, use the '-X' option when opening your ssh connection, e.g. ssh -X vm-hostname Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] [ml-admin] Merge opensuse-cloud and opensuse-virtual
Henne Vogelsang wrote: > Hey Peepz, > > we are currently trying to reduce the amount of communication channels > we provide. Because of this I would like to merge the opensuse-virtual > an opensuse-cloud list. I'm just not sure how. Yeah, not as simple as it initially seems. > Judging from the amount > of traffic we should fold opensuse-cloud into opensuse-virtual and > change the description of -virtual to refer to cloud activities. > Would folks realize opensuse-cloud is the proper forum for their kvm, xen, virtualbox, libvirt, etc. questions? I'm struggling to find a name that represents both cloud and virtualization, which suggests both are needed? Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org To contact the owner, e-mail: opensuse-virtual+ow...@opensuse.org
Re: [opensuse-virtual] [kernel bug + panic] Re: DomU config for a Centos6 guest?
gregk@mailforce.net wrote: > On Tuesday, September 20, 2011 10:56 AM, gregk@mailforce.net wrote: > >> I can't tell you anything further because I can't figure out how to VNC >> from my Desktop to the 'rh6variants' HVM Guest on my headless Dom0 >> server if I can't set an IP address for VNC to listen on, >> > > If in the HVM config file I do both > > - vncpasswd='vncpass' > + #vncpasswd='vncpass' > - vncconsole=10 > + vncdisplay=10 > I'd suggest vnc=1 vncunused=1 vncunused=1 means use first unused port, starting at 5900. This is more flexible than using a static port, since the domain won't start if that port is already in use. vncpasswd should work though. Perhaps there has been a regression. Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org For additional commands, e-mail: opensuse-virtual+h...@opensuse.org
Re: [opensuse-virtual] [kernel bug + panic] Re: DomU config for a Centos6 guest?
gregk@mailforce.net wrote: > Hi Jim, > > On Monday, September 19, 2011 3:03 PM, "Jim Fehlig" > wrote: > >> Config similar to >> >> name="blabla" >> memory=1024 >> vcpus=4 >> ... >> builder="hvm" >> device_model="/usr/lib/xen/bin/qemu-dm" >> kernel="/usr/lib/xen/boot/hvmloader" >> boot="d" >> disk=[ 'file:/var/lib/xen/images/sles11sp1 gmc-hvm/disk0,hda,w', >> 'file:/path/to/centos6.iso,hdc:cdrom,r', ] >> ... >> > > Trying HVM with this config, > > name = 'rh6variants' > builder = 'hvm' > device_model = '/usr/lib/xen/bin/qemu-dm' > kernel = '/usr/lib/xen/boot/hvmloader' > boot = 'd' > disk = [ 'file:/stor/CentOS-6.0-x86_64-netinstall.iso,hdc:cdrom,r', > 'phy:/dev/VG0/Centos6,xvda,w',] > ... > > Gives me this, > Hmm, can you provide the complete config? Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org For additional commands, e-mail: opensuse-virtual+h...@opensuse.org
Re: [opensuse-virtual] DomU config for a Centos6 guest?
gregk@mailforce.net wrote: > Hi Jim, > > On Monday, September 19, 2011 2:43 PM, "Jim Fehlig" > wrote: > >> No bootargs are necessary. The default kernel in domU's >> /boot/grub/menu.lst will be selected (which could be overridden with >> bootargs). >> >> bootargs='' >> >> should suffice. >> > > For an Opensuse 11.4 DomU on Opensuse 11.4 Dom0, using > > cat mail.cfg > name = 'mail' > builder = 'linux' > bootloader = '/usr/lib/xen/boot/domUloader.py' > bootargs = '--entry=xvda1:vmlinuz-xen,/boot/initrd-xen' > ... > > running > > xm create -c mail.cfg > > gives the usual long/verbose output at console, ending with success at > >Welcome to openSUSE 11.4 "Celadon" - Kernel 2.6.37.6-0.7-xen >(xvc0). > >mail login: > > If I switch to using pygrub instead, > > cat mail.cfg > name = 'mail' > builder = 'linux' > bootloader = '/usr/bin/pygrub' > bootargs = '' > bootargs='-q' will stop the pygrub interactive mode and give you the same behavior as domUloader. Forgot to mention that one :). Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org For additional commands, e-mail: opensuse-virtual+h...@opensuse.org
Re: [opensuse-virtual] [kernel bug + panic] Re: DomU config for a Centos6 guest?
gregk@mailforce.net wrote:
> Hi Jim,
>
> On Monday, September 19, 2011 2:08 PM, "Jim Fehlig"
> wrote:
>
>> gregk@mailforce.net wrote:
>>
>>> A buddy suggested I should try a different way, using a Centos
>>> kernel+ramdisk directly instead of trying to dig for it.
>>>
>>>
>> Good suggestion...
>>
>>
>>> mount -o loop CentOS-6.0-x86_64-netinstall.iso /mnt/Centos6
>>> cp -a /mnt/Centos6/isolinux/{vmlinuz,initrd.img} /stor/
>>>
>>>
>> Is that a pv or pv-ops capable kernel?
>>
>
> I wouldn't know how to check for certain. I don't know much about
> Centos6 yet, just that it's supposed to be a clone of RedHat6 minus the
> commercial support.
>
> I did find this from a Citrix forum
>
> Re: Death of multi-os Virtualization?
> http://forums.citrix.com/message.jspa?messageID=1456321
>
> The latest Linux kernels, including the RHEL6 kernel, leverage a
> capability called paravirt_ops -- which means that the kernel is
> enabled right out of the box to work properly in a
> paravirtualized Xen environment (as well as other
> paravirtualization) -- without needing a special "xenified"
> kernel. One of our SEs has taken it and, with a small number of
> simple manual steps, gotten it up and running without a special
> kernel. Red Hat's no longer shipping a Xen hypervisor doesn't
> affect this at all.
>
> So it looks like the CentOS kernel should be a paravirt_ops (same as
> pv-ops I think) kernel.
>
> Shouldn't what I'm doing work with that kernel?
>
>
>> Have you considered installing as an hvm guest?
>>
>
> I'm having a hard time finding Opensuse virtualization "DocsForDummies".
>
> I'd consider anything that works right now. If you can help me tweak my
> config I'll try that.
>
Config similar to
name="blabla"
memory=1024
vcpus=4
...
builder="hvm"
device_model="/usr/lib/xen/bin/qemu-dm"
kernel="/usr/lib/xen/boot/hvmloader"
boot="d"
disk=[ 'file:/var/lib/xen/images/sles11sp1 gmc-hvm/disk0,hda,w',
'file:/path/to/centos6.iso,hdc:cdrom,r', ]
...
/etc/xen/examples/ contains quite a few examples for both pv and hvm.
But vm-install should help with the installation. Just ensure to check
"Full virtualization" for Virtualization Method.
Regards,
Jim
--
To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org
For additional commands, e-mail: opensuse-virtual+h...@opensuse.org
Re: [opensuse-virtual] DomU config for a Centos6 guest?
gregk@mailforce.net wrote: > Hi Jim, > > >>> but I've read that the domUloader.py is an Opensuse-specific bootloader? >>> >>> >> Yes, and not really needed anymore. IIRC, it was originally written as >> a host bootloader for NetWare, which pygrub now supports too. >> > > Oh. So even if booting an Opensuse Guest these days, we should change > > builder = 'linux' > bootloader = '/usr/lib/xen/boot/domUloader.py' > bootargs = '--entry=xvda1:vmlinuz-xen,/boot/initrd-xen' > > to, > > builder = 'linux' > bootloader = '/usr/bin/pygrub' > bootargs = '--entry=xvda1:vmlinuz-xen,/boot/initrd-xen' > No bootargs are necessary. The default kernel in domU's /boot/grub/menu.lst will be selected (which could be overridden with bootargs). bootargs='' should suffice. Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org For additional commands, e-mail: opensuse-virtual+h...@opensuse.org
Re: [opensuse-virtual] DomU config for a Centos6 guest?
gregk@mailforce.net wrote: > Hi, > > I've got no problems launching Opensuse DpmU Guests on top of my > Opensuse Dom0. > [...] > but I've read that the domUloader.py is an Opensuse-specific bootloader? > Yes, and not really needed anymore. IIRC, it was originally written as a host bootloader for NetWare, which pygrub now supports too. Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org For additional commands, e-mail: opensuse-virtual+h...@opensuse.org
Re: [opensuse-virtual] [kernel bug + panic] Re: DomU config for a Centos6 guest?
gregk@mailforce.net wrote:
> A buddy suggested I should try a different way, using a Centos
> kernel+ramdisk directly instead of trying to dig for it.
>
Good suggestion...
> mount -o loop CentOS-6.0-x86_64-netinstall.iso /mnt/Centos6
> cp -a /mnt/Centos6/isolinux/{vmlinuz,initrd.img} /stor/
>
Is that a pv or pv-ops capable kernel?
Have you considered installing as an hvm guest?
Regards,
Jim
> Using this config file
>
> cat /stor/centos6.cfg
> name = 'centos6'
> builder = 'linux'
> kernel = '/stor/vmlinuz'
> ramdisk = '/stor/initrd.img'
> maxmem = 1024
> memory = 1024
> disk = [ 'file:/stor/CentOS-6.0-x86_64-netinstall.iso,hdc:cdrom,r',
> 'phy:/dev/VG0/centos6,xvda,w']
> vif = [ '' ]
> vfb = [ 'type=vnc, vncdisplay=1, vnclisten=127.0.0.1' ]
> extra = 'textmode=1 xencons=xvc0'
> on_crash = 'destroy'
> on_reboot = 'restart'
> on_shutdown = 'destroy'
> localtime = 0
>
> Then 'xmcreate -c /stor/centos6.cfg' crashes at "kernel BUG at
> fs/sysfs/group.c:65!;Kernel panic - not syncing: Fatal exception". The
> console output is below.
>
> I'm stumped for now. I searched online for Centos6, Opensuse, Xen, etc.
> I haven't found a working cfg yet.
>
> Greg
>
> CONSOLE OUTPUT - "xmcreate -c /stor/centos6.cfg"
>
> Using config file "/stor/centos6.cfg".
> Started domain centos6 (id=10)
> Initializing cgroup subsys cpuset
> Initializing cgroup subsys cpu
> Linux version 2.6.32-71.el6.x86_64 (mockbu...@c6b6.centos.org) (gcc
> version 4.4.4 20100726 (Red Hat 4.4.4-13) (GCC) ) #1 SMP Fri May 20
> 03:51:51 BST 2011
> Command line: textmode=1 xencons=xvc0 noirqdebug elevator=noop
> KERNEL supported cpus:
> Intel GenuineIntel
> AMD AuthenticAMD
> Centaur CentaurHauls
> ACPI in unprivileged domain disabled
> BIOS-provided physical RAM map:
> Xen: - 000a (usable)
> Xen: 000a - 0010 (reserved)
> Xen: 0010 - 4000 (usable)
> DMI not present or invalid.
> last_pfn = 0x4 max_arch_pfn = 0x4
> init_memory_mapping: -4000
> RAMDISK: 01cbf000 - 052bf000
> No NUMA configuration found
> Faking a node at -4000
> Bootmem setup node 0 -4000
> NODE_DATA [8000 - 0003bfff]
> bootmap [0003c000 - 00043fff] pages 8
> (7 early reservations) ==> bootmem [00 - 004000]
> #0 [00 - 001000] BIOS data page ==> [00 -
> 001000]
> #1 [00054c2000 - 00054f1000] XEN PAGETABLES ==> [00054c2000 -
> 00054f1000]
> #2 [006000 - 008000] TRAMPOLINE ==> [006000 -
> 008000]
> #3 [000100 - 0001c9eff8]TEXT DATA BSS ==> [000100 -
> 0001c9eff8]
> #4 [0001cbf000 - 00052bf000] RAMDISK ==> [0001cbf000 -
> 00052bf000]
> #5 [00052bf000 - 00054c2000] XEN START INFO ==> [00052bf000 -
> 00054c2000]
> #6 [10 - 2d] PGTABLE ==> [10 -
> 2d]
> Zone PFN ranges:
> DMA 0x0001 -> 0x1000
> DMA320x1000 -> 0x0010
> Normal 0x0010 -> 0x0010
> Movable zone start PFN for each node
> early_node_map[2] active PFN ranges
> 0: 0x0001 -> 0x00a0
> 0: 0x0100 -> 0x0004
> SFI: Simple Firmware Interface v0.7 http://simplefirmware.org
> SMP: Allowing 2 CPUs, 0 hotplug CPUs
> No local APIC present
> APIC: disable apic facility
> PM: Registered nosave memory: 000a - 0010
> Allocating PCI resources starting at 4000 (gap: 4000:c000)
> Booting paravirtualized kernel on Xen
> Xen version: 4.1.2_01-143.1 (preserve-AD)
> NR_CPUS:4096 nr_cpumask_bits:2 nr_cpu_ids:2 nr_node_ids:1
> PERCPU: Embedded 31 pages/cpu @880005527000 s95064 r8192 d23720
> u126976
> pcpu-alloc: s95064 r8192 d23720 u126976 alloc=31*4096
> pcpu-alloc: [0] 0 [0] 1
> Xen: using vcpu_info placement
> Built 1 zonelists in Node order, mobility grouping on. Total pages:
> 257996
> Policy zone: DMA32
> Kernel command line: textmode=1 xencons=xvc0 noirqdebug elevator=noop
> IRQ lockup detection disabled
> PID hash table entries: 4096 (order: 3, 32768 bytes)
> Checking aperture...
> No AGP bridge found
> AMD-Vi disabled by default: pass amd_iommu=on to enable
> Memory: 960752k/1048576k available (4935k kernel code, 388k absent,
> 87436k reserved, 3927k data, 1220k init)
> Hierarchical RCU implementation.
> NR_IRQS:33024 nr_irqs:288
> Console: colour dummy device 80x25
> console [tty0] enabled
> console [hvc0] enabled
> allocated 10485760 bytes of page_cgroup
> please try 'cgroup_disable=memory' option if you don't want memory
> cgroups
> installing Xen timer for CPU 0
> Detected 2806.450 MHz processor.
> Calibrating delay loop (skipped), value calculated using timer
> frequency.. 5612.90 BogoMIPS (lpj=2806450)
> pid_max: default: 32768 minimum: 301
> Security Framework initialized
> SELinux: Initializing.
> Dentry cach
Re: [opensuse-virtual] Re: Guest launches manually OK, not on System reboot when config is in /etc/xen/auto ?
gregk@mailforce.net wrote: > With help in this thread, > http://lists.xensource.com/archives/html/xen-users/2011-09/msg00368.html > > the problem seems to be the 'xl toolstack'. > > Changing in /etc/init.d/xendomains, > > - xm_cmd=xl > + xm_cmd=xm > > Seems to fix the problem. > > Is this a bug in the native xl toolchain, Opensuse's build of it, or > just the script? Or something else? > There could be problems in xl/libxl, and wouldn't surprise me a bit if there are bugs in the script. Ensuring the script works with the new tool stack is on my long list of things to do. As you discovered, using the legacy xm/xend toolstack is a workaround. Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org For additional commands, e-mail: opensuse-virtual+h...@opensuse.org
Re: [opensuse-virtual] Help with KIWI
Athanasios-Ilias Rousinopoulos wrote: > Hello Mates, > > i have installed the "yast2-product-creator" so as to make a live-cd > iso with the KIWI Graphical Tool- Frontends (Image Creator and Product > Creator). > I don't think there are any kiwi developers/experts on this list :-/. You'll probably have better luck with opensuse-factory. Regards, Jim > Well , everything , goes well , but finally i can't build the image > with the repositories and applications that i want. > > Here is the message that appears : > > " Starting SuSEconfig, the SuSE Configuration Tool... > Running module permissions only > Reading /etc/sysconfig and updating the system... > Executing /sbin/conf.d/SuSEconfig.permissions... > Finished. > ls: cannot access /usr/lib*/hal/hald-generate-fdi-cache: No such file > or directory > Jul-14 03:48:28 <3> : Couldn't setup image system failed > Jul-14 03:48:28 <1> : Umounting path: > /tmp/YaST2-07513-SLHE7H/myphysical//var/cache/kiwi > Jul-14 03:48:28 <1> : EXEC [umount > "/tmp/YaST2-07513-SLHE7H/myphysical//var/cache/kiwi" 2>&1] > Jul-14 03:48:28 <1> : Umounting path: > /tmp/YaST2-07513-SLHE7H/myphysical//var/cache/zypp > Jul-14 03:48:28 <1> : EXEC [umount > "/tmp/YaST2-07513-SLHE7H/myphysical//var/cache/zypp" 2>&1] > Jul-14 03:48:28 <1> : Umounting path: /tmp/YaST2-07513-SLHE7H/myphysical/dev > Jul-14 03:48:28 <1> : EXEC [umount > "/tmp/YaST2-07513-SLHE7H/myphysical/dev" 2>&1] > Jul-14 03:48:28 <1> : Umounting path: > /tmp/YaST2-07513-SLHE7H/myphysical/dev/pts > Jul-14 03:48:28 <1> : EXEC [umount > "/tmp/YaST2-07513-SLHE7H/myphysical/dev/pts" 2>&1] > Jul-14 03:48:29 <1> : Umounting path: /tmp/YaST2-07513-SLHE7H/myphysical/sys > Jul-14 03:48:29 <1> : EXEC [umount > "/tmp/YaST2-07513-SLHE7H/myphysical/sys" 2>&1] > Jul-14 03:48:29 <1> : Umounting path: > /tmp/YaST2-07513-SLHE7H/myphysical/var/run/dbus > Jul-14 03:48:29 <1> : EXEC [umount > "/tmp/YaST2-07513-SLHE7H/myphysical/var/run/dbus" 2>&1] > Jul-14 03:48:29 <1> : Umounting path: /tmp/YaST2-07513-SLHE7H/myphysical/dev > Jul-14 03:48:29 <1> : EXEC [umount > "/tmp/YaST2-07513-SLHE7H/myphysical/dev" 2>&1] > Jul-14 03:48:29 <1> : Umounting path: /tmp/YaST2-07513-SLHE7H/myphysical/proc > Jul-14 03:48:29 <1> : EXEC [umount > "/tmp/YaST2-07513-SLHE7H/myphysical/proc" 2>&1] > Jul-14 03:48:29 <3> : KIWI exited with error(s) done > Jul-14 03:48:29 <1> : shmctl RMID: invalid parameter " . > > > I include the whole log files in this e-mail > > What do i have to do so as to solve this problem? > > > Thanks, > > > Rousinopoulos Athanasios-ilias > > > > -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org For additional commands, e-mail: opensuse-virtual+h...@opensuse.org
Re: [opensuse-virtual] Xen live migration and xen-kmp issues
Jan Beulich wrote: > >It seems that HVM emulates a 'PIIX3 IDE' controller and the driver seems > >to be built-in (i.e. not a kernel module): > > Hmm, yes, that's ugly. I have always been advocating for modular > drivers... > > >Things go south when 'xen-platform-pci' or 'xen-vbd' module is loaded... > > > >Is there a way to disable a built-in driver?? > > I'm not aware of one, but we surely had found some way to use PV > drivers in 11.4 guests, just that I don't recall the details - hopefully > others reading the list will. IIRC, use 'hda=noprobe' or similar on guest kernel cmdline to avoid the built-in driver. Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org For additional commands, e-mail: opensuse-virtual+h...@opensuse.org
Re: [opensuse-virtual] Xen4 (11.4) autostart of DomU
Wolfgang Rosenauer wrote: > During testing for getting debugging logs I could confirm that the > on_xend_start option is set correctly if toggled within virt-manager. > But when restarting virt-manager it is shown unchecked as if it wasn't > set to autostart. The virt-manager log doesn't seem to have something > about that issue though. If you'd like to see it anyway please let me know. > I did some testing on an 11.4 system and it seemed to work fine *if* the domain was shutdown. I.e., toggling the autostart option in virt-manager and 'virsh autostart domU [--disable]' worked fine when domU was not running. If domU was running, xend seemed to ignore the request for a change in the autostart setting. So at first glance, this seems to be a bug in xend. Maybe you have time to enter a bug? :-). Cheers, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org For additional commands, e-mail: opensuse-virtual+h...@opensuse.org
Re: [opensuse-virtual] Xen4 (11.4) autostart of DomU
Wolfgang Rosenauer wrote: > Hi, > > I've got a new Xen host based on openSUSE 11.4 but I'm confused on how > to make guests starting automatically when Dom0 is started. > AFAIK, there are 3 options for autostarting domains when the host starts 1. (on_xend_start start) in domU config (as you mentioned) 2. xendomains init script 3. libvirt-guests init script 1) requires xend to successfully start. I assume that is the case? Can you directly edit the domU config with (on_xend_start start) and have it work? > With an older Xen host I was able to do it via virsh or virt-manager > which apparently resulted in (on_xend_start start) in the guest's > config.sxp. > Right. > But on 11.4 the virsh or virt-manager flag has no effect apparently. I > can toggle the flag without an error but revisiting the setting shows it > unchecked again. > Hmm, maybe a regression. If xend is running and autostart works when manually adding to domU config, open a bug and provide /root/.virt-manager/virt-manager.log. virsh output with export LIBVIRT_DEBUG=1 would help tool. Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org For additional commands, e-mail: opensuse-virtual+h...@opensuse.org
Re: [opensuse-virtual] XEN fail after update
Klaus Vink Slott wrote: > I am new to xen and libvirt but I had managed to build a setup on 11.4, > with xen + drbd. I had most things working. Migration back and forth was > functional provided that I used "xm migrate" commands. So this morning I > was quite pleased with my progress so far. > > > But today i ran a zypper up on my setup and now I cant contact xen > anymore. My xend.log after reboot is temporary at > http://klaus.vink-slott.dk/ieshu5Mo.txt > > > I am using the packages from /repositories/Virtualization/openSUSE_11.4/ > > > ss0:~ # rpm -qa | grep xen > > xen-libs-4.1.0_01-106.1.x86_64 > > xen-doc-html-4.1.0_01-106.1.x86_64 > > xen-4.1.0_01-106.1.x86_64 > > xen-tools-4.1.0_01-106.1.x86_64 > Heh, new Xen4.1 packages (with new libxenlight toolstack). We're still shaking out a lot of bugs and issues. I'd stick with Xen 4.0.1 for a while. Regards, Jim > kernel-xen-2.6.37.1-1.2.2.x86_64 > > xen-kmp-desktop-4.1.0_01_k2.6.37.1_1.2-106.1.x86_64 > > ss0:~ # rpm -qa | grep libvirt > > libvirt-python-0.8.8-111.1.x86_64 > > libvirt-0.8.8-111.1.x86_64 > > libvirt-client-0.8.8-111.1.x86_64 > > > My question: Is there anything obvius I should try before downgrading > and waiting for new package? > > > I apologize for my limited English > > -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org For additional commands, e-mail: opensuse-virtual+h...@opensuse.org
Re: [opensuse-virtual] 11.3 on Xen 3.1 (oS 10.3)
Wolfgang Rosenauer wrote: > On 07/27/2010 01:08 AM, Jim Fehlig wrote: > >> Wolfgang Rosenauer wrote: >> >>> Hi, >>> >>> I just tried to install openSUSE 11.3 as guest on a 10.3 (aka Xen 3.1) >>> host but booting the kernel failed immediately. I haven't found useful >>> logfiles but the vm disappears basically immediately again. >>> >>> Should it be possible to run/install 11.3 on Xen 3.1? >>> >> No, the 11.3 xen kernel only has compatibility back to Xen 3.3. >> > > Thanks. That's bad news though. Even more I fear the answer to my next > question then: Would 10.3 guests run on 11.3? > Yes, older guests on new host should work. That's a common use case. Folks want support for their new hardware (upgrade host platform), but stability for their certified workloads (same guest platform). > Upgrading the Dom0 will be a lot of pain but having to upgrade that and > some of the guests at the same time will be almost impossible. > IIRC, the support docs for SLE recommend upgrading the host before upgrading any SLE guests. Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org For additional commands, e-mail: opensuse-virtual+h...@opensuse.org
Re: [opensuse-virtual] 11.3 on Xen 3.1 (oS 10.3)
Wolfgang Rosenauer wrote: > Hi, > > I just tried to install openSUSE 11.3 as guest on a 10.3 (aka Xen 3.1) > host but booting the kernel failed immediately. I haven't found useful > logfiles but the vm disappears basically immediately again. > > Should it be possible to run/install 11.3 on Xen 3.1? > No, the 11.3 xen kernel only has compatibility back to Xen 3.3. Regards, Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org For additional commands, e-mail: opensuse-virtual+h...@opensuse.org
Re: [opensuse-virtual] a 'marketing' question ...
Carlos Ribeiro wrote: > My question is. Microsoft supports these KVM, VirtualBox for their windows > servers? > > IMHO, I believe this would be a good point, especially because today, we have > no more only windows, neither only Linux, in 99% all datacenter are > heteregeneos plataform. If Microsoft not endorses KVM, AFAIK, work is being done to get Windows SVVP certified on KVM. Surely RHEL6 will have SVVP cerfication for important Windows versions. -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org For additional commands, e-mail: opensuse-virtual+h...@opensuse.org
Re: [opensuse-virtual] Question about KVM
Sascha 'saigkill' Manns wrote: > Hello Mates, > > today i've tried to make a new virtual Machine with KVM in YaST. But sadly > i'm getting the Error: "internal error: No > 'bridge' attribute specified with . > Maybe anyone knows what is to do? > What version of the distro are you using? 11.2, 11.3, Factory? -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org For additional commands, e-mail: opensuse-virtual+h...@opensuse.org
Re: [opensuse-virtual] help launching a Centos 5.5 Xen PV guest on an OpenSuse 11.2 Dom0 host ?
Jan Beulich wrote: I have no clue if/when/how synchronization happens between Factory and other repos like the one you use. The Virtualization project in OBS is where we maintain virtualization-related packages like xen, kvm, libvirt, etc. Commits to this project are expected to work reasonably well on Factory. When a package passes "stage" testing, it is submited to Factory. We typically submit packages from Virtualization to Factory at key milestones, such as 11.3 milestone 4, or 11.3 RC2. OBS project admins have control over build targets. For Virtualization project I have all of the currently maintained openSUSE distros enabled. In general, I'd like users to have access to latest version of a package for a maintained distro. Doing so is certainly not supported and bug reports are useful if they exist in Factory as well. For some packages, such as Xen, perhaps it is best to only have Factory target enabled. Jim -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org For additional commands, e-mail: opensuse-virtual+h...@opensuse.org
Re: [opensuse-virtual] help launching a Centos 5.5 Xen PV guest on an OpenSuse 11.2 Dom0 host ?
0bo0 wrote: > On Fri, Jun 4, 2010 at 12:10 AM, Jan Beulich wrote: > >> This certainly isn't with the 11.2 kernel, is it? On the HEAD (Factory) >> one we did such a change, but there still is no need to load the driver >> through /etc/sysconfig/kernel (as that would result in also attempting >> to load the [non-existing] driver on native boots). Instead, the proper >> place is in xend's startup script, as was also done by us in preparation >> to that kernel side change. >> > > heh. i'm getting lost as to what's been (not) done where ... > > *I* currently have > > uname -a > Linux server 2.6.34-42-xen #1 SMP 2010-06-01 15:17:00 +0200 x86_64 > x86_64 x86_64 GNU/Linux > > So you are using the kernel that includes the change, but a xen that doesn't (as mentioned yesterday). The current "HEAD" xen is not building for 11.2 and the repo contains the last good HEAD build of xen on Apr 3, which does not have the change to load evtchn driver. You can add the following to xend init script or build the HEAD xen for 11.2 by disabling building of KMPs. # but for now it's safest to have them loaded when xend starts in dom0. + modprobe evtchn 2>/dev/null || true modprobe blktap 2>/dev/null || true -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org For additional commands, e-mail: opensuse-virtual+h...@opensuse.org
Re: [opensuse-virtual] help launching a Centos 5.5 Xen PV guest on an OpenSuse 11.2 Dom0 host ?
0bo0 wrote: got it. extra= requires "... ksdevice=eth0 method=http://mirrors.kernel.org/centos/5.5/os/x86_64 ..." Nice perseverance :-). Glad to hear you got it sorted out. -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org For additional commands, e-mail: opensuse-virtual+h...@opensuse.org
Re: [opensuse-virtual] help launching a Centos 5.5 Xen PV guest on an OpenSuse 11.2 Dom0 host ?
0bo0 wrote: i've Opensuse 11.2 Xen Dom0, lsb_release -dr Description:openSUSE 11.2 (x86_64) Release:11.2 uname -a Linux server 2.6.34-42-xen #1 SMP 2010-06-01 15:17:00 +0200 x86_64 x86_64 x86_64 GNU/Linux rpm -qa | grep -i xen xen-libs-4.0.0_21087_01-38.1.x86_64 kernel-xen-devel-2.6.34-42.1.x86_64 xen-tools-4.0.0_21087_01-38.1.x86_64 kernel-xen-2.6.34-42.1.x86_64 xen-4.0.0_21087_01-38.1.x86_64 I build & run numerous OpenSuse DomUs -- no issues. I need to netinstall a Centos 5.5 PV'd DomU Guest on this host. Using the procedure & config below, it starts to boot -- then hangs, with not much useful debugging info ... at least, not to my eye, where/how I'm looking. If someone could once-over the following, and make any suggestions/hints as to what's up, it'd be appreciated! Here's the current step-by-step: lvcreate -n vm_boot -L 160M /dev/VG0 lvcreate -n vm_swap -L 512M /dev/VG0 lvcreate -n vm_root -L 8G /dev/VG0 mkdir -p /home/xen/centos55/xenboot/ setenv DL "http://mirrors.kernel.org/centos/5.5"; cd /home/xen/centos55 wget $DL/isos/x86_64/CentOS-5.5-x86_64-netinstall.iso cd ./xenboot/ wget $DL/os/x86_64/images/xen/initrd.img wget $DL/os/x86_64/images/xen/vmlinuz vi /home/xen/centos_install.cfg - name = 'centos' builder = 'linux' kernel= "/home/xen/centos55/xenboot/vmlinuz" ramdisk = "/home/xen/centos55/xenboot/initrd.img" disk = [ 'phy:/dev/VG0/vm_boot,xvda,w', 'phy:/dev/VG0/vm_swap,xvdb,w', 'phy:/dev/VG0/vm_root,xvdc,w', 'file:/home/xen/centos55/CentOS-5.5-x86_64-netinstall.iso,hdc:cdrom,r', ] boot = 'd' vif = [ 'mac=00:16:3E:01:06:00, bridge=brA, vifname=vif.A' ] vfb = [ 'type=vnc, vncdisplay=106, vnclisten=127.0.0.1' ] extra= 'textmode=1 xencons=xvc0 elevator=noop lang=en_US.UTF-8 keyboard=us timezone=America/Los_Angeles bootproto=static ip=192.168.1.155 gateway=192.168.1.1 nameserver=192.168.1.1 method=http://mirrors.kernel.org/centos/5.5/os/x86_64' maxmem = 1024 memory = 1024 vcpus= 4 cpu_cap = 100 cpu_weight = 256 localtime= 0 netif= 'yes' Hmm, I haven't seen the netif option used before. IIUC, it enables the domU to act a network backend - aka driver domain. Does the domU boot progress if this setting is removed? on_shutdown = 'destroy' on_reboot= 'restart' on_crash = 'destroy' - @ xm create -c /home/xen/centos_install.cfg starts ok, but then hangs, - Using config file "/home/xen/centos_install.cfg". Started domain centos (id=5) Bootdata ok (command line is textmode=1 xencons=xvc0 elevator=noop lang=en_US.UTF-8 keyboard=us timezone=America/Los_Angeles bootproto=static ip=192.168.1.155 gateway=192.168.1.1 nameserver=192.168.1.1 method=http://mirrors.kernel.org/centos/5.5/os/x86_64) Linux version 2.6.18-194.el5xen (mockbu...@builder10.centos.org) (gcc version 4.1.2 20080704 (Red Hat 4.1.2-48)) #1 SMP Fri Apr 2 15:34:40 EDT 2010 ... XENBUS: Device with no driver: device/vif/1 XENBUS: Device with no driver: device/vif/2 XENBUS: Device with no driver: device/vif/3 XENBUS: Device with no driver: device/console/0 A pv driver issue in the boot image? Initalizing network drop monitor service Write protecting the kernel read-only data: 483k ( ... HANGS HERE ...) - Not much in the logs. Does 'xm dmesg' or /var/log/messages in host contain anything enlightening? Jim and, tail -f /var/logl/xen/* - ... [2010-06-02 19:46:17 4832] DEBUG (DevController:653) hotplugStatusCallback /local/domain/0/backend/vbd/5/51744/hotplug-status. [2010-06-02 19:46:17 4832] DEBUG (DevController:667) hotplugStatusCallback 1. [2010-06-02 19:46:17 4832] DEBUG (DevController:144) Waiting for 5632. [2010-06-02 19:46:17 4832] DEBUG (DevController:653) hotplugStatusCallback /local/domain/0/backend/vbd/5/5632/hotplug-status. [2010-06-02 19:46:17 4832] DEBUG (DevController:667) hotplugStatusCallback 1. [2010-06-02 19:46:18 4832] DEBUG (DevController:139) Waiting for devices irq. [2010-06-02 19:46:18 4832] DEBUG (DevController:139) Waiting for devices vfb. [2010-06-02 19:46:18 4832] DEBUG (DevController:139) Waiting for devices pci. [2010-06-02 19:46:18 4832] DEBUG (DevController:139) Waiting for devices vusb. [2010-06-02 19:46:18 4832] DEBUG (DevController:139) Waiting for devices vtpm. [2010-06-02 19:46:18 4832] INFO (XendDomain:1207) Domain centos (5) unpaused. ( ... HANGS HERE ...) - -- To unsubscribe, e-mail: opensuse-virtual+unsubscr...@opensuse.org For addition
