On Tue, 2017-03-07 at 19:02 +0000, Fedora compose checker wrote:
Missing expected images:
Atomic qcow2 x86_64
Server dvd i386
Server boot i386
Atomic raw-xz x86_64
Failed openQA tests: 16/107 (x86_64), 1/2 (i386), 1/2 (arm)
New failures (same test did not fail in 26-20170306.n.0):
ID: 60760 Test: x86_64 Server-dvd-iso install_default_upload
URL:
https://openqa.fedoraproject.org/tests/60760
So...I think this one may actually be caused by some hardware issues
we're having with one of the openQA worker host boxes. openQA detected
a system stall during a needle check that failed; it's written to abort
the test any time this happens. But the needle check that failed was
actually *expected* to fail, so if openQA had just carried on the test
would've completed fine.
The stall likely happened because the worker host in question seems to
be periodically overheating and throttling its CPUs, which obviously
causes all the tests to run much slower. We're looking into why this is
happening.
In any case, not a Fedora bug.
ID: 60767 Test: x86_64 Server-dvd-iso
install_repository_nfs_graphical
URL:
https://openqa.fedoraproject.org/tests/60767
This one seems to have hit the 'hang during package install' bug that
sometimes happens lately.
ID: 60809 Test: arm Minimal-raw_xz-raw.xz
install_arm_image_deployment_upload
URL:
https://openqa.fedoraproject.org/tests/60809
This actually fails every day, just sometimes it shows up as 'failed',
sometimes as 'incomplete'. Right now we think it's failing on
https://bugzilla.redhat.com/show_bug.cgi?id=1422634 ; the boot winds up
in emergency mode.
ID: 60821 Test: x86_64 universal install_no_swap@uefi
URL:
https://openqa.fedoraproject.org/tests/60821
Looks like anaconda suddenly crashed in this test. We didn't get a
coredump, though, possibly because I need to update the openQA post-
fail code to handle coredumpctl-handled crashes. Looking in syslog, we
see:
13:42:22,808 INFO kernel:anaconda[1572]: segfault at 0 ip (null) sp
00007ffffd7281f8 error 14 in system-python[55e55d753000+1000]
13:42:22,872 INFO systemd:Started Process Core Dump (PID 1931/UID 0).
13:42:27,803 CRIT systemd-coredump:Process 1572 (anaconda) of user 0 dumped
core.#012#012Stack trace of thread 1572:#012#0 0x0000000000000000 n/a (n/a)
Installed system changes in test x86_64 universal
install_package_set_kde:
Mount / contents changed to 149.7529804% of previous size
System load changed from 1.57 to 1.84
Previous test data:
https://openqa.fedoraproject.org/tests/60113#downloads
Current test data:
https://openqa.fedoraproject.org/tests/60811#downloads
The sudden increase in size here seems like an interesting result that
could bear looking into. I'll try and find a minute to do it if I can,
but anyone else could too, by installing the KDE package set from the
20170306.n.0 compose tree and the 20170307.n.0 compose tree, comparing
the results to verify the same thing happens, then trying to see what
eats the space...
--
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | XMPP: adamw AT happyassassin . net
http://www.happyassassin.net