Urgent F23 testing request: logout bugs

Giulio 'juliuxpigface' juliux.pigface at gmail.com
Wed Oct 21 19:28:37 UTC 2015


On wed, 21/10/2015 at 09.13 -0700, Adam Williamson wrote:
> Hi folks!
> 
> So we're building 23 Final RC2 now, but we have a couple of
> outstanding
> proposed blocker bugs, and we could use some more data to evaluate
> them.
> 
> The way to test is really simple:
> 
> * Install F23 Final TC11 Workstation or KDE live[1]
> * Create a user and log in
> * Log out a bunch of times. Just keep logging in and out.
> 
> Ideally try this both in a VM and on a real system, if you have a
> spare
> system to test on.
> 
> There are two bugs that you might hit:
> 
> 1. https://bugzilla.redhat.com/show_bug.cgi?id=1272737
> 2. https://bugzilla.redhat.com/show_bug.cgi?id=1273247
> 
> With the first one, the system gets stuck at a black screen after
> logging out. You won't be able to switch to a TTY or anything, but if
> you have another system, you'll find that ssh'ing in still works. So
> far it seems this bug *only* affects Workstation (we believe it
> involves Wayland, and Workstation has GDM-on-Wayland). It seems easy
> to
> reproduce in a KVM, harder to reproduce on a real system, but some
> testers have reproduced it on a real system after several attempts.
> We're particularly interested in whether you hit it on a real system,
> and if so, how often.
> 
> With the second bug, assuming you don't hit the first bug, the cursor
> disappears on the login screen, possibly after you click somewhere
> (e.g. the password text entry box). Roshi says he sees something like
> this on Workstation but not KDE, kparal says he sees it on KDE but
> not
> Workstation, and I can't reproduce it on either. This one is probably
> only reproducible in a KVM, but of course let us know if you manage
> to
> trigger it on a real system!
> 
> If you could report back to the list what configuration you tested
> on,
> and whether and how often you managed to reproduce either bug, that
> would be a great help. For virtual machine testing, please let us
> know:
> 
> * What virtualization system you used (KVM/virt-manager is preferred)
> * For a KVM:
>   ** Whether your VM has a virtual 'tablet' input device (one
>      is often created by default, so check in virt-manager)
>   ** What video adapter and viewer you're using (e.g. qxl/SPICE or
>      std/VNC)
>   ** What OS and release is running on the host system
>   ** The graphics adapter and driver being used on the host system
> 
> Thanks!
> 
> [1] Downloads:
>     https://dl.fedoraproject.org/pub/alt/stage/23_TC11/Workstation/i3
> 86/iso/Fedora-Live-Workstation-i686-23-TC11.iso
>     https://dl.fedoraproject.org/pub/alt/stage/23_TC11/Workstation/x8
> 6_64/iso/Fedora-Live-Workstation-x86_64-23-TC11.iso
>     https://dl.fedoraproject.org/pub/alt/stage/23_TC11/Live/i386/Fedo
> ra-Live-KDE-i686-23-TC11.iso
>     https://dl.fedoraproject.org/pub/alt/stage/23_TC11/Live/x86_64/Fe
> dora-Live-KDE-x86_64-23-TC11.iso
> -- 
> Adam Williamson
> Fedora QA Community Monkey
> IRC: adamw | Twitter: AdamW_Fedora | XMPP: adamw AT happyassassin .
> net
> http://www.happyassassin.net
> 
> 

I've tested the "mouse cursor" bug (the one reported by Roshi) with
~15/20 attempts: I haven't been able to reproduce it.

Here are some details:
1) qemu-kvm via virt-manager on a Fedora 22 host
2) Yes, my VM has got that virtual 'tablet', created by default.
3) The video adapter is QXL/SPICE here.
4) Two Amd GPUs (4250m+5650m) with radeon driver on the host.
5) KDE x86_64 TC11 on the guest.

// Giulio (jpigface)


More information about the test mailing list