Graphical nstallation of Beta TC4 Fedora-Workstation-netinst-x86_64-22_Beta_TC4.iso fails in VirtualBox

[Adam Williamson]

On Sun, 2015-03-22 at 09:14 -0700, Adam Williamson wrote:
> On Sun, 2015-03-22 at 08:44 -0700, Adam Williamson wrote:
> > On Sun, 2015-03-22 at 09:30 +0100, Joachim Backes wrote:
> > > For testing Fedora-Workstation-netinst-x86_64-22_Beta_TC4 I set 
> > > up  a  virtual machine in VirtualBox with a virtual 32MB video 
> > > card.  After  the  installer has started and set up the X, 
> > > rather early  it hangs  up.  Inspecting the logs I saw a hint: 
> > > please install  F22 in 
> > > *textmode*.  Trying that will install completely.
> > > 
> > > After installation, the F22 VM runs flawlessly in a X11 
> > > environment.
> > > 
> > > Question: is this a bug or a feature?
> > 
> > A bug, but bugs that are VBox-specific don't necessarily rate a  
> > super  high priority. What exactly is this message you're seeing 
> > in  the logs?
> 
> Interestingly, I see the same bug in my openQA tests:
> 
> https://openqa.happyassassin.net/tests/2782
> 
> screen goes black shortly after reaching the hub. I don't see this  
> message about text mode, though.
> 
> I'm wondering if it's something to do with this time jump:
> 
> 15:42:59,993 DEBUG anaconda: spoke is ready: 
> <pyanaconda.ui.gui.spokes.storage.StorageSpoke object at 
> 0x7ff3615dea50>
> 09:55:00,003 INFO anaconda: System time set to Sun Mar 22 16:55:00 
> 2015 UTC
> 09:55:00,000 INFO anaconda: System time set to Sun Mar 22 16:55:00 
> 2015 UTC
> 09:55:00,000 INFO anaconda: System time set to Sun Mar 22 16:55:00 
> 2015 UTC
> 09:55:00,000 INFO anaconda: Thread Done: AnaDateTimeThread 
> (140683642418944)
> 09:55:00,006 DEBUG anaconda: spoke is ready: 
> <pyanaconda.ui.gui.spokes.datetime_spoke.DatetimeSpoke object at 
> 0x7ff361670a10>
> 
> in my openQA jobs the screen seems to blank right around the time 
> that  happens (though with the accelerated videos it's hard to be 
> sure).

Nope, in a manual test, the blank happens about 25 seconds after the 
time jump. (The manual test VM has Spice/QXL, while my openQA uses 
VNC/vga, so this doesn't seem tied to a specific video adapter or to 
VNC...)

In journalctl -b I see a couple of interesting messages right around 
the blank:

Mar 22 17:41:33 localhost org.a11y.Bus[1477]: 
g_dbus_connection_real_closed: Remote peer vanished with error: 
Underlying GIOStream returned 0 bytes on an async read (g-io-error-
quark, 0). Exiting.
Mar 22 17:41:33 localhost org.a11y.atspi.Registry[1483]: 
g_dbus_connection_real_closed: Remote peer vanished with error: 
Underlying GIOStream returned 0 bytes on an async read (g-io-error-
quark, 0). Exiting.

but haven't been able to figure much out beyond that yet...

TC4 and TC2 (which definitely doesn't have this bug for me) have the 
same Xorg versions, but TC4 has kernel rc3.git0.1.fc22 while TC2 has 
kernel rc1.git0.1.fc22.
-- 
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | XMPP: adamw AT happyassassin . net
http://www.happyassassin.net