QA » openQA Project » openQA Tests

Richard the links become broken. Have you new one?

By the way i tryed to run clonejob.pl for ppc64 and it wasn't working. Is there a way to simulate ppc jobs locally? or should i have a ppc machine? thx

dmaiocchi wrote:

Richard the links become broken. Have you new one?

any job for migration_offline_sle12_ppc or migration_offline_sle12_allpatterns_ppc on any build is a good example :)

https://openqa.suse.de/tests/280729
https://openqa.suse.de/tests/280730

By the way i tryed to run clonejob.pl for ppc64 and it wasn't working. Is there a way to simulate ppc jobs locally? or should i have a ppc machine? thx

In THEORY you should be able to get qemu to run qemu-ppc with a ppc64le CPU without needing a ppc machine...never tried it..but with a suitably creative vars.json you should be able to make isotovideo do it on an intel machine

we did do it for aarch64, with the aarch64-emu machine, with the following vars:

BIOS=qemu-uefi-aarch64.bin
CDMODEL=virtio-blk-device
HDDMODEL=virtio-blk-device
QEMU=aarch64
QEMUCPU=cortex-a57
QEMUMACHINE=virt
QEMU_NO_KVM=1
TIMEOUT_SCALE=4
WORKER_CLASS=qemu_x86_64

Something similar for power should work, IN THEORY

ok i will try this out

ok isn't that fixed richard? i see yast_snapper test.

what do you mean by missing snapper? missing a snapper test, or missing the package itself?

ok i got, i cannot find any image in openqa.suse.de that match GA.

is it ok if i use this image?

SLE-12-Server-ppc64le-GM-allpatterns.qcow2

will provide disk image to you...

done:
wget http://10.162.2.8/sles12.gm.all-patterns.brtfs.snapper.qcow2

zluo, the image you provided is only 10GB and fails when installing packages during upgrade https://openqa.suse.de/tests/463859#step/install_and_reboot/4, probably because out of disk space.

I am more or less wandering blindly here but called the following to create a SLE12SP0 installation. maybe this is creating what we want:

openqa_client_osd jobs post DISTRI=sle VERSION=12 FLAVOR=Server-DVD ARCH=ppc64le BUILD=okurz_poo9714 \
ISO=SLE-12-Server-DVD-ppc64le-GM-DVD1.iso INSTALLONLY=1 QEMU_COMPRESS_QCOW2=1 \
PUBLISH_HDD_1=SLES-12-GM-gnome-ppc64le_snapper_20g.qcow2 TEST=create_gm_ppc_image \
MACHINE=ppc64le WORKER_CLASS=qemu_ppc64le HDDSIZEGB=20 MAX_JOB_TIME=86400 TIMEOUT_SCALE=10

--> https://openqa.suse.de/tests/463876

Failed with "cirrus" as VGA not available. Triggered with

openqa_client_osd jobs post DISTRI=sle VERSION=12 FLAVOR=Server-DVD ARCH=ppc64le BACKEND=qemu \
NOVIDEO=1 OFW=1 QEMUCPU=host SERIALDEV=hvc0 BUILD=okurz_poo9714 \
ISO=SLE-12-Server-DVD-ppc64le-GM-DVD1.iso INSTALLONLY=1 QEMU_COMPRESS_QCOW2=1 \
PUBLISH_HDD_1=SLES-12-GM-gnome-ppc64le_snapper_20g.qcow2 TEST=create_gm_ppc_image \
MACHINE=ppc64le WORKER_CLASS=qemu_ppc64le HDDSIZEGB=20 MAX_JOB_TIME=86400 TIMEOUT_SCALE=10

--> t#463883

EDIT: Completed now. Let's see what happens if I clone sle-12-SP2-Server-DVD-ppc64le-Build1649-migration_offline_sle12_ppc@ppc64le but with the new HDD_1

openqa_clone_job_osd 462022 HDD_1=SLES-12-GM-gnome-ppc64le_snapper_20g.qcow2

--> t#463890

Let's see if it can find snapshots or if the harddisk of the SP0 image is still too small or snapshots won't be enabled by default at all.

EDIT: t#463890 failed with timeout on first boot. I suspect either a sporadic performance problem or network configuration problem. retriggered with TIMEOUT_SCALE

openqa_clone_job_osd 463890 TIMEOUT_SCALE=10

--> t#463947

EDIT: cleaned temporary builds

$ openqa_client_osd jobs/463859 delete
{ result => 1 }
$ openqa_client_osd jobs/463724 delete
{ result => 1 }

Result of t#463947 is still that it can't boot into a desktop session although switching to terminal and logging in works ok. I checked logs but couldn't find anything obvious. Anyone else has an idea?

[08/07/2016 11:21:03] <okurz> [08/07/2016 11:06:13] <okurz> How wants to take a look at https://openqa.suse.de/tests/464628#live (30 minutes left) and tell me how I can further debug the bootup process? It's stuck there
[08/07/2016 11:32:17] <okurz> ok, fine. any further ideas how to follow on? I would think switching to text console and evaluate why display manager does not show up. the logs show that gdm is running at least
[08/07/2016 12:49:09] <okurz> at least we know now better that "vncviewer" is easier to use to also switch terminals and such
[08/07/2016 13:03:34] <okurz> that is what my test did! It's just not booting but I actually checked in 464665 that there are snapshots now. I will reconduct it and try to record a video

openqa_clone_job_osd 464665 NOVIDEO=0

then connected to instance when it stalled with

vncviewer -Shared malbec.arch:91

changed to text console (press F8 in vncviewer, select ctrl and alt in menue, exit menue, press F2), login as root and called some commands for debugging. As it looks, at least there are snapshots now as visible in the video
https://openqa.suse.de/tests/464666/file/video.ogv at 1:47.

So, what now?

used my experiences from here to write https://progress.opensuse.org/projects/openqatests/wiki/Wiki#Tips-for-test-development-and-issue-investigation

Make sure not to loose what we already observed with the failing test cases to support Feature 317784: Implement a check and warning to make sure we don't start a Service Pack migration with insufficient disk space

Investigated https://openqa.suse.de/tests/464665/file/first_boot-journal.log with the help of rbrown. What I see is

Jul 08 06:15:01 susetest systemd[1]: Started User Manager for UID 0.
Jul 08 06:15:24 susetest dbus[1455]: [system] Activating service name='org.opensuse.Snapper' (using servicehelper)
Jul 08 06:15:24 susetest systemd[1]: Starting Cleanup of Temporary Directories...
Jul 08 06:15:24 susetest dbus[1455]: [system] Successfully activated service 'org.opensuse.Snapper'
Jul 08 06:15:24 susetest systemd[1]: Started Cleanup of Temporary Directories.
Jul 08 06:16:05 susetest kernel: BTRFS info (device vda3): relocating block group 9156165632 flags 1
Jul 08 06:16:05 susetest kernel: BTRFS info (device vda3): found 1960 extents
Jul 08 06:16:06 susetest kernel: BTRFS info (device vda3): found 1958 extents
Jul 08 06:16:06 susetest kernel: BTRFS info (device vda3): relocating block group 10364125184 flags 1
…

Our idea is that there is some btrfs cleanup pending which is taking so long that systemd is terminating the wait and preventing a proper X startup? I will try to either boot into single mode first and see if it settles down before first boot or give the vanilla GA session more time before saving the image (after some btrfs cleanup grace period)

I scheduled poo_9714 test in osd:
medium: sle-12-Server-DVD-ppc64le
group: Test Development: SLE 12 SP2
parameters:
DESKTOP=gnome
HDDSIZEGB=20
INSTALLONLY=1
PUBLISH_HDD_1=poo_9714.qcow2

hi mkravec, did you read the previous comments? What did you change so that your installation could yield a usable system?

I read them - nothing changed, I just have to start somewhere..

I reproduced https://openqa.suse.de/tests/510844#step/first_boot/3 on ppc VM. Does not look like openqa issue.
I compared working 10GB image we currently have with the one I produced, packages and autoinst.xml are the same.

https://bugzilla.suse.com/show_bug.cgi?id=993253

SLES-12 needs to be patched before upgrade to SP2, otherwise there are issues mentioned above.

wow, you achieved what I failed to do.

Thank for providing instructions.

Just for the record: How did you run the virtual machine? qemu-ppc64le locally or on a ppc64le worker?

Regarding the "low disk space" warning: There is https://progress.opensuse.org/issues/11438 to implement an actual test for this. For now we have the test implicitly because they fail because of out of diskspace. What about just preserving the "too small" images for test_suites which have that in mind, e.g.

• use the new, bigger drive for test_suite migration_offline_sle12_allpatterns
• create a new test_suite migration_offline_sle12_allpatterns_too_low_space which uses the old, too small disk

does this make sense?

makes sense to me..can we do it?

great work Martin!