os-autoinst-distri-opensuse
os-autoinst-distri-opensuse copied to clipboard
DESKTOP=gnome System's ram must be at least 4GB
https://progress.opensuse.org/issues/153808
The failed tests are expected and No such issue with memory size=4g https://openqa.suse.de/tests/13352892#step/first_boot/10
Great PR! Please pay attention to the following items before merging:
Files matching lib/**.pm
:
- [ ] Consider adding or extending unit tests in t/
This is an automatically generated QA checklist based on modified files.
I like the concept and the code is perfect, but why do we need yet another variable to pass just for this specific case ? e.g. GNOME_MEMORY_GE_4G=1
?
I like the concept and the code is perfect, but why do we need yet another variable to pass just for this specific case ? e.g.
GNOME_MEMORY_GE_4G=1
?
From my comment on the progress ticket:
we will still need to add a followup where we certify that effectively 2gb is the absolute minimum, with kdump disabled for gnome
That variable basically will pave the way for the follow up
I like the concept and the code is perfect, but why do we need yet another variable to pass just for this specific case ? e.g.
GNOME_MEMORY_GE_4G=1
?From my comment on the progress ticket:
we will still need to add a followup where we certify that effectively 2gb is the absolute minimum, with kdump disabled for gnome
That variable basically will pave the way for the follow up
Thanks for your kindly comments, I wanted to use NO_KDUMP
as a switch, but seems it might not be a perfect one, since we may install the system without kdump but enable kdump tests later :)
And add a new variable GNOME_MEMORY_GE_4G=1
can help us to eliminate effects for some workers with limit memory resource [xen/vmware etc]
@foursixnine If you want to approve/merge this PR, please help inform other squads, since it will introduce new failures if QEMURAM<4096 :)
@foursixnine If you want to approve/merge this PR, please help inform other squads, since it will introduce new failures if QEMURAM<4096 :)
They have been informed via Slack and via Matrix https://suse.slack.com/archives/C02CANHLANP/p1706264121953869 but we can mention them here:
@os-autoinst/tests-maintainer FYI
@rfan1 can you check on Tumbleweed and Leap and adapt changes if necessary? @DimStar77 @lkocman
@foursixnine If you want to approve/merge this PR, please help inform other squads, since it will introduce new failures if QEMURAM<4096 :)
They have been informed via Slack and via Matrix https://suse.slack.com/archives/C02CANHLANP/p1706264121953869 but we can mention them here:
@os-autoinst/tests-maintainer FYI
@rfan1 can you check on Tumbleweed and Leap and adapt changes if necessary? @DimStar77 @lkocman
https://en.opensuse.org/Hardware_requirements_15.5, I can see that no such requirement for leap, but for TW, I failed to find the document
@foursixnine If you want to approve/merge this PR, please help inform other squads, since it will introduce new failures if QEMURAM<4096 :)
They have been informed via Slack and via Matrix https://suse.slack.com/archives/C02CANHLANP/p1706264121953869 but we can mention them here: @os-autoinst/tests-maintainer FYI @rfan1 can you check on Tumbleweed and Leap and adapt changes if necessary? @DimStar77 @lkocman
https://en.opensuse.org/Hardware_requirements_15.5, I can see that no such requirement for leap, but for TW, I failed to find the document
I'm pretty sure Leap and TW will have the same limitation, so we'd need to check by running some tests, but I would like to avoid a is_opensuse in the memory check in general
@foursixnine If you want to approve/merge this PR, please help inform other squads, since it will introduce new failures if QEMURAM<4096 :)
They have been informed via Slack and via Matrix https://suse.slack.com/archives/C02CANHLANP/p1706264121953869 but we can mention them here: @os-autoinst/tests-maintainer FYI @rfan1 can you check on Tumbleweed and Leap and adapt changes if necessary? @DimStar77 @lkocman
https://en.opensuse.org/Hardware_requirements_15.5, I can see that no such requirement for leap, but for TW, I failed to find the document
I'm pretty sure Leap and TW will have the same limitation, so we'd need to check by running some tests, but I would like to avoid a is_opensuse in the memory check in general
Well, even is_tumbleweed
is not recommended, this new logic is more simple :)
Most GNOME tests on TW use QEMURAM=1536 and work fine.
The linked bug report is specific to ppc64le and some SAP stuff FWICT.
Most GNOME tests on TW use QEMURAM=1536 and work fine.
The linked bug report is specific to ppc64le and some SAP stuff FWICT.
We'll follow up https://progress.opensuse.org/issues/153808#note-6 to find out the lower limit and normalize it for both, while the bug was specific for ppc64le, we've had the same problem with other architectures, specially when enabling kdump (which takes considerably larger portion on ppc64le) and likely create the follow up ticket during review
FWICT this will break pretty much all GNOME tests on Leap 15.5, 15.6 and Tumbleweed?
FWICT this will break pretty much all GNOME tests on Leap 15.5, 15.6 and Tumbleweed?
This we need to modify before merging, so that openSUSE is not affected, and is aligned with the strategy.
as it is Friday... next week perhaps :D - The minimum memory requirements will be tested via: https://github.com/os-autoinst/os-autoinst-distri-opensuse/pull/18521#issuecomment-1920807145 as a follow up
FWICT this will break pretty much all GNOME tests on Leap 15.5, 15.6 and Tumbleweed?
This we need to modify before merging, so that openSUSE is not affected, and is aligned with the strategy.
as it is Friday... next week perhaps :D - The minimum memory requirements will be tested via: #18521 (comment) as a follow up
Draft PR for o3, https://github.com/os-autoinst/opensuse-jobgroups/pull/424.
For SLE tests, It should be better if each squad can take care of it's own job groups :)
Mark it as draft now, since we still need check with other squads and O3 user to adapt the job settings before merging this PR. I will double confirm with @foursixnine on this ticket.