F19: replacing chardev patches with upstream ones
by Amit Shah
Hello,
Fedora has carried flow control patches for qemu chardevs since a very
long time now. They weren't upstream, and they were required for
spice, usb-redir to work properly and not freeze the guest.
Upstream qemu has now merged an alternative implementation of the same
concept, and most of the bugs have been shaken out upstream (we don't
know of any known bugs now). So I'm of the view we should replace
the older patches in F19's qemu with the backport of the newer,
upstream ones.
Is this fine? I can work with Cole / Hans who have been keeping the
older patches uptodate on Fedora to replace the patches.
Thanks,
Amit
8 years, 7 months
Fedora virt status April 2013
by Cole Robinson
Here are some (slightly early) virt status bits for April. Fairly quiet month
with no major new releases of qemu or libvirt.
Fedora 19
=========
The F19 alpha was released on April 23rd! The important bits:
https://fedoraproject.org/get-prerelease
https://fedoraproject.org/wiki/F19_Alpha_release_announcement
https://fedoraproject.org/wiki/Common_F19_bugs
The release slipped one week due to some installation issues on UEFI machines,
but the general consensus is that things are working pretty well so far.
The current schedule is:
- May 28: Beta Release
- May 28: Virt Test Day
- May 30: Spice Test Day
- Jun 18: Final change deadline
- Jul 02: GA release
F19 Virt Test Day
=================
The Fedora 19 Virt Test Day is May 28th. Unfortunately this is the same day as
the beta release, and any more slippage means we will be before the beta. But
since the alpha seems reasonably solid so far, we should still have decent
bits to test on. Expect more noise here and on the landing page as we get closer:
https://fedoraproject.org/wiki/Test_Day:2013-05-28_Virtualization
Two days after the virt test day is a spice specific test day. So plan on
sticking around for that one as well.
Bug stats
=========
Bug count on Apr 02 2013 : 244
Total bugs : 260
Not surprisingly as F19 gears up there's been a lot of bugs filed, which is
where most of the growth has been.
By release:
* Fedora 17 : 82
* Fedora 18 : 126
* Fedora 19 : 38
* Fedora rawhide : 14
By package:
* bochs : 1
* gnome-boxes : 29
* gtk-vnc : 1
* ipxe : 4
* libcacard : 1
* libguestfs : 7
* libosinfo : 7
* libseccomp : 1
* libvirt : 47
* libvirt-cim : 2
* libvirt-glib : 2
* libvirt-qmf : 1
* libvirt-sandbox : 18
* netcf : 5
* python-virtinst : 3
* qemu : 55
* spice-gtk : 3
* spice-vdagent : 2
* usbredir : 1
* virt-dmesg : 2
* virt-manager : 21
* virt-top : 1
* virt-v2v : 3
* virt-viewer : 6
* virt-what : 2
* virtio-win : 4
* xen : 8
* xorg-x11-drv-cirrus: 4
* xorg-x11-drv-qxl : 19
By status:
* ASSIGNED : 22
* NEW : 229
* ON_QA : 3
* POST : 6
Thanks,
Cole
8 years, 7 months
Host Graphical Console
by Dean Hunter
On Tue, 2013-04-09 at 17:23 -0500, Dean Hunter wrote:
> On Mon, 2013-04-08 at 16:30 -0400, Cole Robinson wrote:
> > On 04/06/2013 11:05 AM, Dean Hunter wrote:
> > > Could virt-manager be enhanced to open a graphical console on a remote
> > > VM host similar to the way it can open a graphical console on a remote
> > > VM guest?
> > >
> >
> > No I don't think that's anything we will ever add to virt-manager. I can
> > understand the desire but virt-manager is not about managing physical hosts.
> >
> I do not so much want to manage the host as to use scripts to rebuild
> the guests before each round of testing. Sometimes the scripts need a
> little editing to add or remove workarounds for problems between builds.
>
> I really like the feedback provided by the cpu, disk and network
> activity graphs in virt-manager. I can start a rebuild and get a visual
> indication when it has completed. Then I double-click to open a session
> to a VM.
>
> The rest of the GUI was very helpful while I was learning about libvirt
> etc., but GUI's are slow and error prone when I need to repeat something
> several times, hence the use of scripts. I was hoping for the best of
> both worlds.
>
> > > Today I have to use "ssh -X" or xrdp to the remote VM host. "ssh -X"
> > > seems to leave something running if I open gedit for a quick
> > > modification to a file as a logout needs to be followed by a ctrl-C to
> > > get a command prompt. And xrdp seems to have a number of issues, besides
> > > being "yet another" user interface.
> > >
> >
> > There's also vnc-server/xvnc and xspice. But if your main qualm with ssh -X is
> > that the ssh session doesn't exit cleanly after running gedit, I'd a) file a
> > gedit bug and b) learn a simple command line text editor like nano to work
> > around the issue in the interim.
> >
> Yes, I have tried vnc-server/xvnc and xspice. xvnc vs xrdp is an
> interesting debate and depends on the client I expect to need to use.
> The last time I looked at xspice it was not ready to run as a service;
> see https://bugzilla.redhat.com/show_bug.cgi?id=885881.
>
> > Thanks,
> > Cole
>
> I do appreciate your response. It confirms that I have not overlooked a
> significant alternative. I will report the bugs in ssh and Remote
> Desktop Viewer.
>
>
After further review, the problem I am experiencing with Gnome products
from the SSH command line is a long standing design issue with access to
dbus and auto-starting dbus-launch. See
https://bugzilla.redhat.com/show_bug.cgi?id=663467
So I guess I need a directory of scripts that can be shared between the
VM host and its peer machine that I use for most personal activities
like music, e-mail, browsing from bookmarks, etc.; executing scripts
from an SSH command line on the VM host and editing the scripts from the
Gnome desktop on the personal machine. Would you use NFS to share the
directory between two Fedora machines or is there another alternative?
8 years, 7 months
bridge networking busted?
by Tom Horsley
I got the latest updates (as far as I know), running
virt-manager-0.9.5-1.fc18.noarch
When I tried to install a minimal fedora 18 virtual
machine, virt-manager said "no network".
I've got 2 different bridges on my host, and it
wouldn't offer either one.
I had to install the machine and edit the xml
later to add a bridged network device.
Is this a known bug? Is it likely to get fixed
sometime? It is pretty irritating.
8 years, 7 months
F19a virt-manager experiences
by Reartes Guillermo
Hello,
I recently subscribed to this list.
I am testing F19a with KVM/Virt-Manager
* ulquiorra is a F17 (updated) host (main worksation)
* stark is a F19a RC4 (updated) Host, Virtualization is installed here.
I am aware that it is not expected to work yet, i am just reported the
issues i have found.
If i connect from ulquiorra to stark via 'ssh -Y' the remote virt-manager
closes
itself when changing anaconda screens with the following in virt-manager's
logs:
[Sat, 20 Apr 2013 17:04:17 virt-manager 2435] DEBUG (cli:83) Uncaught
exception:
Traceback (most recent call last):
File "/usr/share/virt-manager/virtManager/console.py", line 518, in
_channel_new_cb
self.audio = spice.Audio(self.spice_session)
RuntimeError: could not create SpiceAudio object
So i tried virt-viewer to see if it makes a difference:
# virt-viewer 3 &
[1] 8950
(virt-viewer:8950): GSpice-WARNING **: PulseAudio context failed Connection
refused
(virt-viewer:8950): GSpice-WARNING **: pa_context_connect() failed:
Connection refused
(virt-viewer:8950): GSpice-WARNING **: Warning no automount-inhibiting
implementation available
(/usr/bin/virt-viewer:8950): Spice-CRITICAL **:
../spice-common/common/canvas_base.c:802:canvas_get_lz: condition
`(image->descriptor.type == SPICE_IMAGE_TYPE_LZ_PLT) || (n_comp_pixels ==
width * height)' failed
Thread 5 (Thread 0x7fa79da37700 (LWP 8964)):
#0 0x0000003cf14ea8fd in poll () from /lib64/libc.so.6
#1 0x00007fa7a7d36234 in g_main_context_iterate.isra.22 () from
/lib64/libglib-2.0.so.0
#2 0x00007fa7a7d3633c in g_main_context_iteration () from
/lib64/libglib-2.0.so.0
#3 0x00007fa7a7d36389 in glib_worker_main () from /lib64/libglib-2.0.so.0
#4 0x00007fa7a7d5a2c5 in g_thread_proxy () from /lib64/libglib-2.0.so.0
#5 0x0000003cf1c07c53 in start_thread () from /lib64/libpthread.so.0
#6 0x0000003cf14f4ecd in clone () from /lib64/libc.so.6
Thread 4 (Thread 0x7fa79c8d0700 (LWP 8966)):
#0 0x0000003cf14ea8fd in poll () from /lib64/libc.so.6
#1 0x00007fa7a7d36234 in g_main_context_iterate.isra.22 () from
/lib64/libglib-2.0.so.0
#2 0x00007fa7a7d3633c in g_main_context_iteration () from
/lib64/libglib-2.0.so.0
#3 0x00007fa79c8d89ad in dconf_gdbus_worker_thread () from
/usr/lib64/gio/modules/libdconfsettings.so
#4 0x00007fa7a7d5a2c5 in g_thread_proxy () from /lib64/libglib-2.0.so.0
#5 0x0000003cf1c07c53 in start_thread () from /lib64/libpthread.so.0
#6 0x0000003cf14f4ecd in clone () from /lib64/libc.so.6
Thread 3 (Thread 0x7fa78e652700 (LWP 8968)):
#0 0x0000003cf1c0b922 in pthread_cond_timedwait@(a)GLIBC_2.3.2 () from
/lib64/libpthread.so.0
#1 0x00007fa7a7d75705 in g_cond_wait_until () from /lib64/libglib-2.0.so.0
#2 0x00007fa7a7d0ca81 in g_async_queue_pop_intern_unlocked () from
/lib64/libglib-2.0.so.0
#3 0x00007fa7a7d0d00b in g_async_queue_timeout_pop () from
/lib64/libglib-2.0.so.0
#4 0x00007fa7a7d5ad56 in g_thread_pool_thread_proxy () from
/lib64/libglib-2.0.so.0
#5 0x00007fa7a7d5a2c5 in g_thread_proxy () from /lib64/libglib-2.0.so.0
#6 0x0000003cf1c07c53 in start_thread () from /lib64/libpthread.so.0
#7 0x0000003cf14f4ecd in clone () from /lib64/libc.so.6
Thread 2 (Thread 0x7fa78de51700 (LWP 8969)):
#0 0x0000003cf14ea8fd in poll () from /lib64/libc.so.6
#1 0x00007fa7a7d36234 in g_main_context_iterate.isra.22 () from
/lib64/libglib-2.0.so.0
#2 0x00007fa7a7d3669a in g_main_loop_run () from /lib64/libglib-2.0.so.0
#3 0x00007fa7a8555b86 in gdbus_shared_thread_func () from
/lib64/libgio-2.0.so.0
#4 0x00007fa7a7d5a2c5 in g_thread_proxy () from /lib64/libglib-2.0.so.0
#5 0x0000003cf1c07c53 in start_thread () from /lib64/libpthread.so.0
#6 0x0000003cf14f4ecd in clone () from /lib64/libc.so.6
Thread 1 (Thread 0x7fa7a44c7a80 (LWP 8950)):
#0 0x0000003cf1c0e0cd in read () from /lib64/libpthread.so.0
#1 0x0000003ea6244261 in spice_backtrace_gstack () from
/lib64/libspice-client-glib-2.0.so.8
#2 0x0000003ea624ba07 in spice_logv () from
/lib64/libspice-client-glib-2.0.so.8
#3 0x0000003ea624bb58 in spice_log () from
/lib64/libspice-client-glib-2.0.so.8
#4 0x0000003ea623cd81 in canvas_get_lz.constprop.24 () from
/lib64/libspice-client-glib-2.0.so.8
#5 0x0000003ea623d3c7 in canvas_get_image_internal () from
/lib64/libspice-client-glib-2.0.so.8
#6 0x0000003ea623eead in canvas_draw_copy () from
/lib64/libspice-client-glib-2.0.so.8
#7 0x0000003ea6224d27 in display_handle_draw_copy () from
/lib64/libspice-client-glib-2.0.so.8
#8 0x0000003ea621b532 in spice_channel_recv_msg () from
/lib64/libspice-client-glib-2.0.so.8
#9 0x0000003ea621cda0 in spice_channel_coroutine () from
/lib64/libspice-client-glib-2.0.so.8
#10 0x0000003ea62435fb in coroutine_trampoline () from
/lib64/libspice-client-glib-2.0.so.8
#11 0x0000003ea6243476 in continuation_trampoline () from
/lib64/libspice-client-glib-2.0.so.8
#12 0x0000003cf14473a0 in ?? () from /lib64/libc.so.6
#13 0x0000000002298480 in ?? ()
#14 0x0000000000000000 in ?? ()
I was able to use virt-manager remotelly from ulquiorra before the last
massive stark update, it was slow (expected, since both host and guest run
f19a with debugging enabled) but this worked.
Cheers.
8 years, 7 months
Fedora 19
by Dean Hunter
When using virt-install to build a Fedora 19 VM guest on a Fedora 18 VM
host, what is the correct value for the "--os-variant" parameter as
"fedora19" does not seem to be available?
8 years, 7 months
virsh output discrepancy with vol-list
by Shergill, Gurinder
Hello,
I am seeing following discrepancy with virsh output for volume list, which goes away when I restart the libvirtd daemon.
Here virsh shows 4 volumes. Also, notice that first 2 volumes are duplicates.
$ virsh vol-list --pool default
Name Path
-----------------------------------------
crab.vol.0.img /var/lib/libvirt/images/crab.vol.0.img
crab.vol.0.img /var/lib/libvirt/images/crab.vol.0.img
fish.vol.0.img /var/lib/libvirt/images/fish.vol.0.img
fish.vol.1.img /var/lib/libvirt/images/fish.vol.1.img
However, none of the above volumes currently exist (as I had just deleted them using, again, using virsh).
$ ll /var/lib/libvirt/images/
total 0
Now, I restart libvirt. And, after that virsh output is fine.
$ virsh vol-list --pool default
Name Path
-----------------------------------------
Any idea what is going on here ? Is this a known issue ?
I am using Fedora 17.
Sunny
8 years, 7 months
Host Graphical Console
by Dean Hunter
Could virt-manager be enhanced to open a graphical console on a remote
VM host similar to the way it can open a graphical console on a remote
VM guest?
Today I have to use "ssh -X" or xrdp to the remote VM host. "ssh -X"
seems to leave something running if I open gedit for a quick
modification to a file as a logout needs to be followed by a ctrl-C to
get a command prompt. And xrdp seems to have a number of issues, besides
being "yet another" user interface.
8 years, 7 months
Make virt-viewer go away?
by Tom Horsley
I'm trying to switch to virt-viewer since spicec is
no longer supported, and I find it monumentally
irritating that when I shutdown the virtual machine
virt-viewer won't just go away when the machine is
gone, but instead insists on popping up a dialog
that says "Hey! The machine when away!" (I paraphrase :-)
which I am then forced to click "OK" on.
Is there some cryptic option I can set to make it just
go away and not bother me with stuff I already know?
8 years, 7 months
