It seems Rawhide auto-signing is working fine now: openQA claims
this needle has 'never' been matched (which really means 'not
for a long time'), and I can't find any test fail in the last
year which looks like it landed on this 'authenticate' screen
but the needle failed to match, or anything like that.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
https://openqa.stg.fedoraproject.org/tests/424393 is a failure
where the 'Download' [updates] button was already visible when
we went to the tab. We already checked whether an 'apply' button
is visible and skipped the 'refresh' click if so, but because
the 'download' button is a new thing, we weren't skipping the
'refresh' click if 'download' was already visible.
So in this case, even though we could already see 'download', we
went ahead and clicked 'refresh'...then *immediately* started
looking for 'download'. It seems that Software did not refresh
and remove the 'Download' button *immediately* when we pressed
'refresh' - it left the 'Download' button visible briefly, and
*in this brief window*, we clicked it. *Then* Software kinda
'noticed' we'd clicked 'Update', and it seems it just sort of
throws away our click on 'Download' at that point and does the
refresh.
So at that point, the test thinks it's clicked 'Download' and
expects to see 'Apply', but actually the 'Download' click got
more or less thrown away, so the test fails, sitting at the
'Download' button.
To solve this, let's just extend the existing check to skip the
'refresh' click if 'download' *or* 'apply' are already visible.
There is a sort of possibility here that we could wind up
downloading and installing some updates that existed and were
noticed *before* we did our python3-kickstart trick, but not
install the python3-kickstart update, and cause the test to fail
because of that, but that doesn't seem to have happened before
when we were seeing the 'update' button, so I think I'm not
going to borrow trouble. If it happens, we'll deal with it I
guess.
The comment talks only about KDE, but clearly it can be the case
that an automatic check makes the button visible on GNOME too,
so let's rewrite the comment too.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
The fix for this bug was sent to all releases now, so we should
not need the workaround any more. Let's kill it.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
This is breaking the memory_check tests. I just reproduced it
manually and the UI *does* come back to life if you wait some
time; let's see if we can work around the bug this way.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
If an update test fails before reaching advisory_post, we don't
generate the 'what update packages were installed' and 'were
any update packages *not* installed when they should have been'
logs, but these may well be useful for diagnosing the failure -
so let's also do the same stuff there. Only let's not do it all
twice.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
We hit an interesting case in update testing recently:
https://bodhi.fedoraproject.org/updates/FEDORA-2018-115068f60e
An earlier version of that update failed testing. When we dug
into it a bit, we found that the test was failing because an
earlier version of the `pki-server` package was installed than
the version that was in the update; when asked (as part of
FreeIPA deployment) to install it, dnf had noticed that there
were dependency issues with the version of the package from the
update, but it happened to be able to install the version from
the frozen 'stable' repo...so it just went ahead and did that.
In this case, the 'missed' package resulted in a test failure,
but it'd actually be possible for this to happen and the test
to complete; we really ought to notice when this happens, and
treat it as a test failure.
So what this attempts to do is: at the end of all update tests,
check for all installed packages with the same name as a package
from the update, and compare their full NEVR to the one of the
package from the update. If a package with the same name as one
of the update packages is installed, but does not appear to be
the *same NEVR*, we fail, and upload the lists of packages for
manual investigation as to what the heck's going on.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
From local experimentation, it still actually produces the
output, even though it prints the message about the order being
wrong and exits 1.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
Weirdly, occasionally some update tests seem to fail because
the 'comm' util we use to produce the list of packages from the
update that were actually tested during the job doesn't think
one of the input files is in alphabetical order, even though we
sort them both when they're produced. I don't know if this is
possibly due to the definition of 'alphabetical order' changing
as part of the update, or what. But we really shouldn't *fail*
the test when this happens, as it's not part of the functional
test, we're just producing convenience data. So, let's handle
the command failing, and if it happens, upload the input files
so we can maybe figure out why it's unhappy...
Signed-off-by: Adam Williamson <awilliam@redhat.com>
The previous commit would lead to the 'workaround' getting hit
incorrectly, and might have had some other issues...tweak it a
bit.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
GNOME Software 3.30.5 split the offline update process into two
separate 'download' and 'apply' phases. So we need to handle
clicking 'download' before 'apply', if that happens.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
Somehow, recently, FreeIPA tests are running into Firefox not
quitting because it's showing a warning about closing multiple
tabs. (I think we didn't *get* multiple tabs before but now we
do, for some reason). So let's work around this by clicking
"Close tabs" if the warning appears.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
Lately, we can't be sure the desktop will be on tty1 after we
do 'systemctl isolate graphical.target'. For recent Workstation
lives it actually shows up on tty2.
We could be 'clever' and switch to tty2 on F29+ Workstation
lives...but actually it seems like if we just don't do anything,
systemd switches us to the correct tty. So let's rely on that,
at least as long as it's working.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
At least one test (desktop_notifications_postinstall) boots from
the disk image uploaded by install_default_upload, and needs to
access the grub menu. On F29+ Workstation this is failing,
because the grub menu is now hidden by default, so when the test
boots, it never sees the bootloader screen, and fails.
I considered trying to teach it to hold down shift or hit f8 or
esc at the right time, but that seems like it might be hard. So
instead let's just try to disable the hidden menu when we're
about to upload the installed system image. This is kinda going
against the 'preserve natural system behaviour' principle we try
to use for openQA, but I think it's OK as we do have other tests
that will exercise the 'hidden boot menu' stuff to some extent.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
On F27 we don't get a 'Software is up to date' screen because
there's an upgrade available. Let's work with the refresh button
instead.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
We're not seeing *exactly* #1314991 any more, but we're seeing
something that looks quite similar: the first attempt to find
updates just doesn't find any. No error message, no updates. I
have reported a bug for this and am investigating it, in the
meantime, let's restore the workaround, elaborated a bit, and
looking for the 'Software is up to date' screen instead of the
error message.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
I rather suspect the *bug* is still basically present and it's
why this test often fails, but we no longer seem to see the
*error message* which lets us detect the bug happening. This
needle has not been hit by any test for six months. So let's
remove the workaround as it adds complexity.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
For some reason, in recent tests, switching to a console after
live install completes is taking a long time, and tests are
failing because we 'only' allow 10 seconds for the login prompt
to appear. This seems to indicate some kind of performance bug,
but we don't really want all liveinst tests to fail on in, this
is not primarily a performance testing framework. So let's
tweak the root_console / console_login bits a bit to allow a
configurable timeout for the login prompt to appear, and use
that to wait 30 secs instead of 10 in this case.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
The FreeIPA upgrade test didn't actually check that FreeIPA is
actually running after the upgrade and reboot, it just kinda
assumed it is. Let's add a check to the start of the 'check'
test module that makes sure ipa.service actually comes up to
'active' state. This'll make it clearer when tests are failing
because FreeIPA didn't come up right after the upgrade. The
check will run on non-upgrade tests too, but that's fine.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
This adds a new test intended to just check boot chain things
for updates. It doesn't run any test modules besides the stock
update ones, but sets a variable, ADVISORY_BOOT_TEST, which
causes _advisory_update to do some additional stuff after
installing the updates but before rebooting: it forces regen
of the initramfs and bootloader config, and reinstalls the
bootloader on BIOS (not UEFI as it's not relevant). If the
following boot fails, we probably have a bug somewhere.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
There's this annoying problem where the screen sometimes goes
messed up after ipa-server-uninstall. 'clear' doesn't seem to
really work to fix it up either. Let's try flipping between
ttys. I don't like this much as it's already a pain trying to
work out / remember what tty we might possibly be on at any
given time, but I think we're always on either 1 or 3 here, so
let's do ctrl-alt-f1 ctrl-alt-f3 to ensure at least one change
and wind up on tty3...
Signed-off-by: Adam Williamson <awilliam@redhat.com>
Looking at this, it's a bit weird: the updated packages are
actually included in the upgrade process, but we still run
_advisory_update, which does basically nothing...then reboots.
That's kinda silly and makes the tests a bit flaky, let's fix
it. I don't think there's actually any problem with doing the
upload of updatepkgs.txt in _repo_setup_updates, becase that
already guards against being run more than once, it just bails
very early if it's already been run.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
I'm going to figure out if it's a bug that it takes so long, but
for now let's just bump the timeout.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
RHBZ#1625572 is for gnome-initial-setup running in 'first login'
mode after it's already run in 'user creation' mode (which isn't
meant to happen). This works around that so the subsequent tests
can run. We don't soft-fail because meh.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
If USER_LOGIN is false we can just return; when we reach the
login screen. We don't need a huge conditional when we don't do
anything *after* it, in the false case.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
It transpires that decommissioning wipes some stuff, like the
dirsrv logs. Obviously we want these included in the logs we
upload for reference purposes, so let's upload earlier.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
This adds a set of jobs to test FreeIPA replication. We deploy
a server, deploy a replica of that server, then enrol a client
against the replica and run the client tests.
At first I was planning to add the replica testing into the
main set of FreeIPA tests, but the test ordering/blocking (via
mutexes and barriers and what-have-you) just turns into a big
nightmare that way. This way seems rather simpler to deal with.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
We'd really like to know if FreeIPA is working aside from this
crasher bug, so let's workaround it.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
We kind of want to know if FreeIPA is working aside from this
known bug, so let's treat it as a soft failure and work around
it. But only for Rawhide, not for F27/F28 updates tests.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
https://bugzilla.redhat.com/show_bug.cgi?id=1600823 shows a
case where systemd throws a service that would usually have been
started out of the boot process *entirely* in order to resolve a
dependency loop. This means the service won't show up as failed,
it will just be inactive when it should be active. This still
should constitute a failure of this test, so let's add a check
for the log message that indicates this situation.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
