1
0
mirror of https://pagure.io/fedora-qa/os-autoinst-distri-fedora.git synced 2024-12-05 03:23:09 +00:00
Commit Graph

2530 Commits

Author SHA1 Message Date
Adam Williamson
375575cd07 Use a second disk for the update repo in image build tests
This tweaks the update live and installer image build tests to
store the update repo on a separate disk. This should hopefully
avoid the tests failing due to insufficient disk space when the
update is huge (e.g. KDE or GNOME megaupdates).

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-03-13 12:57:57 -07:00
Adam Williamson
126109f8ac D'oh, missed a quote mark
Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-03-13 11:25:58 -07:00
Adam Williamson
4c281e3d23 Set new PARALLEL_CANCEL_WHOLE_CLUSTER for all parallel parents
This is the magic variable to disable openQA's "cancel whole
cluster on child failure" behaviour that I'm adding in
https://github.com/os-autoinst/openQA/pull/2017 . We want it
set for all parallel parent suites in Fedora.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-03-13 11:20:15 -07:00
Michel Normand
ff3768b1d0 grub2-install parm set to /dev/vda1 for PowerPC
Signed-off-by: Michel Normand <normand@linux.vnet.ibm.com>
2019-03-12 19:41:20 +00:00
Michel Normand
58d6dbb5e6 Add missing advisory_boot ppc64le in templates-updates
and reorder ppc64 and ppc64le for updates_server* flavor
to ease comparison with x86_64.

Signed-off-by: Michel Normand <normand@linux.vnet.ibm.com>
2019-03-12 19:41:20 +00:00
Michel Normand
1b546261c5 Add PowerPC support for tests upgrade of FreeIPA
This is using the two tests previously added by
"Test upgrade of FreeIPA server and client deployment"

Signed-off-by: Michel Normand <normand@linux.vnet.ibm.com>
2019-03-12 19:41:20 +00:00
Adam Williamson
6dde0ea0c9 Fix #1674045 workaround for bootloader case
The initial implementation here has a problem if we spot the
'successful' screen briefly, then the system reboots normally,
reaches the bootloader and proceeds past it all within 10
seconds; in this case we'll never actually spot the bootloader
and do our stuff. This tweak should continue through the code
block immediately if the bootloader shows up during the ten
seconds, otherwise check again for the 'successful' screen and
reboot.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-03-11 18:34:13 -07:00
Adam Williamson
2e75cfe5db Add new variant Firefox certificate error needle
It looks different in Fedora 30 update tests for some reason.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-03-08 07:47:33 -08:00
Adam Williamson
b777094b7b Give live image build command another 10 minutes
40 seems not long enough, it's timing out while the build is
still running sometimes.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-03-06 19:53:31 -08:00
Adam Williamson
d6c339e9ff Disable the 'f30-build' repo for the desktop update test
...GNOME Software doesn't like it.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-03-06 19:51:09 -08:00
Adam Williamson
bb8d9d26d7 Workaround f30 compose/buildroot mismatch
The f30 tag is currently a lot fresher than what's in the repos,
because we haven't had a compose for five days. This causes
problems in some F30 update tests because the update works fine
with what's in the `f30` tag but not what's on the mirrors. To
deal with this until we get a successful compose, let's just
enable the f30-build repo in the update test process...

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-03-06 10:41:16 -08:00
Adam Williamson
b35492e8b5 Add variant GNOME Software refresh needle
desktop_update is failing on F30 update tests because this icon
looks a bit different...

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-03-05 08:13:20 -08:00
Adam Williamson
0c63065ebd Add the #1674045 workaround to one more place
Sigh, so much copypasta makes me sad, should've factored it out.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-03-05 07:35:11 -08:00
Adam Williamson
86fe913b84 Soft fail workaround the 'upgrade hangs at end' bug (#1674045)
We want to get accurate results from the rest of the upgrade
test and we know about this bug, so let's make it a soft fail
so we can see how the test is functioning otherwise.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-03-04 18:07:04 -08:00
Adam Williamson
f3caa1fc53 Bump timeout for install_lang_selected to appear a bit
It seems 3 secs was a bit tight for recent Branched and Rawhide,
test was failing when the screen was just a bit slow to update
for some reason.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-03-04 17:32:25 -08:00
Adam Williamson
450f843389 desktop_browser: try both button and ctrl-t for opening new tab
Sigh. It seems sometimes the button isn't visible because GNOME
pops up a fricking notification that blocks it.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-03-01 09:13:54 -08:00
Adam Williamson
f89e08e16d Add some new needle variants that failed ppc64 tests
Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-03-01 08:52:25 -08:00
Adam Williamson
03a56eb0da Avoid spurious repo check error for universal tests on netinst
As there is *always* a base repo error when booting a netinst
image, this check was giving us spurious failures when running
universal tests on a netinst (which we are currently doing on
ppc64 it seems). Tweak it a bit to hopefully avoid this.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-03-01 08:49:08 -08:00
Adam Williamson
6f24fd452c Seems we need a 15 sec sleep for the Firefox cert workaround
Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-02-26 12:05:55 -08:00
Adam Williamson
b23dd7e14c Use correct name for file to disable Firefox shield studies
Dunno where I got distribution.ini from...

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-02-26 10:57:26 -08:00
Adam Williamson
072333a0f7 Update needle for apps 'System Monitor'
Icon changed.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-02-25 11:30:19 -08:00
Adam Williamson
855e54d8df Refresh workaround for occasional FreeIPA web UI cert issue
Investigation suggests that the Firefox certificate validation
failure we sometimes get when loading the FreeIPA web UI in the
freeipa_webui test module is a race issue in Firefox itself:

https://bugzilla.mozilla.org/show_bug.cgi?id=1530429

refreshing the page seems to work around the problem, so let's
do that.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-02-25 10:13:33 -08:00
Adam Williamson
965722e6b3 Add new variant 'Getting Started' needle
The 'largetext' version of the needle is needed when there's
some notification displaying over the window title, it seems
F28's current font rendering doesn't match the current version
of that needle...

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-02-22 08:40:50 -08:00
Adam Williamson
945a93aa05 Get more info on failures caused by Firefox certificate issues
We're getting an intermittent case where FreeIPA tests fail
because of a web server certificate issue. Click 'Advanced' in
Firefox when this happens so we can get a bit more info on the
problem.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-02-21 12:16:42 -08:00
Adam Williamson
8f0108176b D'oh, move disable_firefox_studies up in utils
...so things that appear earlier in the file can use it.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-02-21 10:35:51 -08:00
Adam Williamson
3b236c7a39 Try and disable Firefox 'shield studies'
I've seen a few tests fail lately because Firefox suddenly
opened a tab with some "Privacy Study" in it and switched to it.
Per https://bugzilla.mozilla.org/show_bug.cgi?id=1529626 , this
should be the way to disable these...

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-02-21 09:51:07 -08:00
Lukas Ruzicka
690788f107 Reneedle the test suite
This is another take of needles to make the suite work again.
2019-02-21 12:29:36 +01:00
Lukas Ruzicka
1e8834825d Add new set of needles (first run)
These needles fix the test suite after the looks has been changed.
2019-02-20 15:52:38 +01:00
Adam Williamson
fc1c8185f0 Add updated graphical encryption passphrase needle
The screen appearance changed in Rawhide, new needle needed.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-02-15 12:10:36 -08:00
Adam Williamson
c1a85c0d72 Add updated grey KDE network notification needle
Same font change as blue.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-02-14 12:35:53 -08:00
Adam Williamson
6d2031dbfd desktop_browser: fix wait_still_screen, drop wait_idle
wait_idle was deprecated and effectively it just sleeps, so
replace these wait_idles with explicit sleeps - I can't think
of a better choice here as we just need to wait for Firefox to
calm down after various operations.

The wait_still_screens were always hitting the 30 second time
out because of the flashing cursor; setting the similarity level
required to lower than the default 47 should fix this.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-02-13 09:51:34 -08:00
Adam Williamson
965a1a6546 Click instead of using ctrl-t for new browser tab
ctrl-t seems to be failing often in Firefox tests, and I can't
figure out why. Let's try clicking the 'new tab' button instead
as a workaround.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-02-13 09:35:32 -08:00
Adam Williamson
1bbb62af32 Drop GDM no logo needle as the logo is back in 0212.n.1...
Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-02-13 00:19:47 -08:00
Adam Williamson
a2a51ca41c Add various new needle variants for Rawhide changes
A bunch of stuff, probably mostly more GTK+ icon theme changes.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-02-13 00:15:50 -08:00
Adam Williamson
42fb8491a2 Re-do the GDM no logo needle
Using translated text was a poor choice...

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-02-11 16:29:30 -08:00
Adam Williamson
9cae91cd89 Add updated needles for Adwaita icon theme changes
The Adwaita icon theme got some changes in a git snapshot build
for Rawhide recently. These changed icons break various needles,
we need new ones.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-02-11 13:12:13 -08:00
Adam Williamson
200755814f Add variant GDM needle for missing Fedora logo
Fedora logo is missing from GDM in today's Rawhide. Not going
all the way and making this a soft fail as the bug is known and
should be fixed soon, I'll take the needle out again then.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-02-11 13:10:46 -08:00
Adam Williamson
0b5dd623aa Add updated KDE blue network notification needle variant
Text display changed here.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-02-11 13:03:15 -08:00
Adam Williamson
809f291a1f Add a couple of needles needed for live image tests (F28)
Forgot to include these in the live image test commit.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-02-08 08:35:35 -08:00
Adam Williamson
f50f6d66b1 Fix up the 'skip update on support server' stuff (a20ea59)
The previous attempt at this (in a20ea59) was wrong, because it
meant we didn't create the update repo at all, and we do need
the repo to exist in the support_server test every time (as the
point of the support server is to serve it out via NFS). We just
don't want to *use it on the support_server system itself* if
the releases don't match. So rejig things to do that properly.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-02-08 08:20:24 -08:00
Adam Williamson
99302c6fd4 Add a live image build test for updates
Just like the installer image build test, only...it builds a live
image. This involves reimplementing quite a chunk of the Koji
livemedia task. Ah, well. Also involves rethinking the flavor
names a bit here, these seem...better.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-02-07 18:28:24 -08:00
Adam Williamson
a20ea59961 Don't install updates on support server if versions don't match
I caught a few failed support_server update tests which failed
because packages from the update weren't installed; they weren't
installed because the update was for F28, but the support server
runs on an F29 disk image. Obviously we should skip the 'install
packages from the update' step in the support_server test unless
the versions match.

Arguably we could *never* install updates in support_server, but
I think it's actually valuable to know if the update breaks
support_server if we *are* on the right version.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-02-07 10:37:04 -08:00
Adam Williamson
578a254cca Add GTK+ 3.24.5 desktop update needle
Didn't notice it in the big retake.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-02-07 10:12:27 -08:00
Adam Williamson
5cbc7df810 Work around GGO gnome-software/582 till it's fixed
We're getting a spurious "Authentication required" for a
background refresh of the system repos sometimes, depending on
whether it happens while we're at a tty. This is now accepted
by upstream, so let's handle it and make it a soft failure for
now.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-02-07 09:58:10 -08:00
Adam Williamson
bd4edb0da7 GTK+ 3.24.5 needle retake
New GTK+ changed something so the background of many interface
elements is a slightly lighter grey, this broke a bunch of
needles. Here are the retakes. Includes one not-strictly-related
SDDM update and a rename of the one @lruzicka did to match the
others.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-02-06 13:27:30 -08:00
Lukas Ruzicka
b182edf697 Add new needles to make tests work
OpenQA tests were choking on _boot_to_anaconda, because
the rawhide_accept_fate was not recognized. I created
a new needle to make it work.
2019-02-05 15:46:22 +01:00
Adam Williamson
8a68b04a3b Run the install_default_update test on UEFI as well
Obviously an installer image may work fine for a BIOS install
but fail for a UEFI install, so we should test it both ways.
This requires the invention of a new (terribly named, but I
couldn't think of anything better...) variable to allow us to
use the 'run test on machine A after test run on machine B'
mechanism without just hardcoding '64bit' as the machine name
in the START_AFTER setting (which would break if we ever wanted
to run update tests on any *other* arch).

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-01-29 15:49:59 +01:00
Adam Williamson
8d1d150798 Handle running update tests on Koji tasks
We quite often want to run the update tests on a Koji task (not
a Bodhi update) for some reason - usually to test a potential
fix for an issue, or at a maintainer's request to test a change
before it is merged upstream and officially sent out as an
update. Up till now I've always hacked up utils.pm on the
staging server by hand to do this, which is horrible. Together
with a commit to fedora_openqa, this should allow us to do it in
a nice, sane way via the CLI. It's mostly just tweaking the
"updates" repo setup in utils.pm as you'd expect, but there's a
bit of subtlety to it because of the installer tests that use
%ADVISORY% as a variable substitution in the disk image name;
you can't do something like `%ADVISORY or KOJITASK%`, sadly, so
I had to have almost-redundant variables ADVISORY, KOJITASK and
ADVISORY_OR_TASK (we could kinda just live with ADVISORY_OR_TASK
except I didn't want to drop ADVISORY as it's an unnecessary
change from previous behavior).

Signed-off-by: Adam Williamson <awilliam@redhat.com>
2019-01-29 11:54:15 +01:00
Lukas Ruzicka
be3aadc249 Change the target of the needle for Clocks in StartStop test
The needle could sometimes fail because the item it was looking
for only appeared in a limited set of situations. The new
location should be safer to check and the test should not
fail as a result.
2019-01-29 10:48:28 +01:00
Lukas Ruzicka
83ff7a0679 Delete Evolution needles.
Evolution has been removed from the default installation of Workstation
so the test is no more needed.
2019-01-28 14:54:58 +01:00