After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 751696 - Virtual machines sometimes don't show up on in the collection
Virtual machines sometimes don't show up on in the collection
Status: RESOLVED NOTGNOME
Product: gnome-boxes
Classification: Applications
Component: general
unspecified
Other Linux
: Normal normal
: --
Assigned To: GNOME Boxes maintainer(s)
GNOME Boxes maintainer(s)
: 754095 755987 757412 760129 764500 767802 (view as bug list)
Depends on:
Blocks:
 
 
Reported: 2015-06-30 07:09 UTC by Adrien Plazas
Modified: 2016-09-20 08:15 UTC
See Also:
GNOME target: ---
GNOME version: ---



Description Adrien Plazas 2015-06-30 07:09:39 UTC
Sometime when starting Boxes, virtual machines don't appear in the collection until I restart the application. This warning is thrown when it happen:
Boxes-WARNING **: libvirt-broker.vala:105: Impossible d'ouvrir qemu+unix:///session: Cannot write data: Noeud final de transport n'est pas connecté

The french part roughly translate to "Final transport node isn't connected".
Comment 1 Zeeshan Ali 2015-06-30 12:34:40 UTC
(In reply to Adrien Plazas from comment #0)
> Sometime when starting Boxes, virtual machines don't appear in the
> collection until I restart the application. This warning is thrown when it
> happen:
> Boxes-WARNING **: libvirt-broker.vala:105: Impossible d'ouvrir
> qemu+unix:///session: Cannot write data: Noeud final de transport n'est pas
> connecté
> 
> The french part roughly translate to "Final transport node isn't connected".

Translation of translation isn't always a good idea. :) You might want to run with en_US locale or check the fr.po file to see the actual error.


Anyway, sounds like some issue with connecting to libvirt. If I had to guess the autostart of libvirt takes too long and connection timesout.
Comment 2 Zeeshan Ali 2015-09-01 21:40:42 UTC
I have seen that happening too recently (especially when I'm not connected to the Internet). IIRC the error message is: Transport endpoint is not connected.

Daniel, do you know this one?
Comment 3 Zeeshan Ali 2015-09-01 21:42:46 UTC
BTW, when this happens, it seems like libvirt gets hang on autolaunch cause if I you immediately restart Boxes, the error is gone. i-e it only happens (if it does) connecting to libvirt when it's not running already.
Comment 4 Zeeshan Ali 2015-09-01 22:09:11 UTC
*** Bug 754095 has been marked as a duplicate of this bug. ***
Comment 5 Zeeshan Ali 2015-09-01 22:10:08 UTC
jimmac is actually able to reproduce this 100% of the time and he filed a bug on libvirt: https://bugzilla.redhat.com/show_bug.cgi?id=1259070 .
Comment 6 Michael Catanzaro 2015-10-02 18:46:41 UTC
*** Bug 755987 has been marked as a duplicate of this bug. ***
Comment 7 Michael Catanzaro 2015-11-05 22:41:47 UTC
Red Hat #1259070 has been closed, but this is still broken with gnome-boxes-3.18.1-1.fc23, libvirt-daemon-1.2.18.1-2.fc23, libvirt-glib-0.2.2-1.fc23.
Comment 8 Michael Catanzaro 2015-11-05 22:45:08 UTC
By the way, the English error is:

(gnome-boxes:18856): Boxes-WARNING **: libvirt-broker.vala:98: Unable to open qemu+unix:///session: Cannot write data: Transport endpoint is not connected

It happens 100% for me the first time I start Boxes after logging in to my desktop, or after having not used Boxes in a long time. It can always be fixed by simply closing Boxes and reopening it. I guess starting Boxes triggers some action that makes the VMs available (starting the virtual network thing?), but Boxes doesn't wait for it to finish before giving up.
Comment 9 Zeeshan Ali 2015-11-12 18:02:47 UTC
(In reply to Michael Catanzaro from comment #7)
> Red Hat #1259070 has been closed, but this is still broken with
> gnome-boxes-3.18.1-1.fc23, libvirt-daemon-1.2.18.1-2.fc23,
> libvirt-glib-0.2.2-1.fc23.

That doesn't mean it's a Boxes issue. :)

(In reply to Michael Catanzaro from comment #8)
>
> It happens 100% for me the first time I start Boxes after logging in to my
> desktop, or after having not used Boxes in a long time. It can always be
> fixed by simply closing Boxes and reopening it. I guess starting Boxes
> triggers some action that makes the VMs available (starting the virtual
> network thing?), but Boxes doesn't wait for it to finish before giving up.

Try reproducing against virsh and virt-manager (you'd need to tell it to connect to session libvirt) and if you can't reproduce with those, then it's a Boxes issue.
Comment 10 Zeeshan Ali 2015-11-13 14:51:10 UTC
*** Bug 757412 has been marked as a duplicate of this bug. ***
Comment 11 Michael Catanzaro 2015-12-16 15:27:00 UTC
(In reply to Zeeshan Ali (Khattak) from comment #9)
> Try reproducing against virsh and virt-manager (you'd need to tell it to
> connect to session libvirt) and if you can't reproduce with those, then it's
> a Boxes issue.

virsh # list
error: failed to connect to the hypervisor
error: no valid connection
error: Cannot write data: Transport endpoint is not connected
Comment 12 Zeeshan Ali 2016-01-05 14:02:40 UTC
*** Bug 760129 has been marked as a duplicate of this bug. ***
Comment 13 Debarshi Ray 2016-01-05 14:17:44 UTC
(In reply to Zeeshan Ali (Khattak) from comment #9)
> (In reply to Michael Catanzaro from comment #7)
> > Red Hat #1259070 has been closed, but this is still broken with
> > gnome-boxes-3.18.1-1.fc23, libvirt-daemon-1.2.18.1-2.fc23,
> > libvirt-glib-0.2.2-1.fc23.
> 
> That doesn't mean it's a Boxes issue. :)

Has anyone filed a new libvirt bug? Or shall I file one?
Comment 14 Michael Catanzaro 2016-01-05 14:24:35 UTC
https://bugzilla.redhat.com/show_bug.cgi?id=1271183

libvirt developers don't seem interested: "Someone is going to need to instrument the code to figure out where the problem is."
Comment 15 Zeeshan Ali 2016-01-05 14:51:40 UTC
(In reply to Michael Catanzaro from comment #14)
> https://bugzilla.redhat.com/show_bug.cgi?id=1271183
> 
> libvirt developers don't seem interested: "Someone is going to need to
> instrument the code to figure out where the problem is."

Cole is more a virt-manager dev than libvirt and I think he is talking to libvirt devs in there mainly.
Comment 16 Zeeshan Ali 2016-04-05 10:50:31 UTC
*** Bug 764500 has been marked as a duplicate of this bug. ***
Comment 17 Zeeshan Ali 2016-06-22 21:48:16 UTC
*** Bug 767802 has been marked as a duplicate of this bug. ***