QA (public) » openQA Project (public) » openQA Tests (public)

I have a feeling this is related to https://bugzilla.opensuse.org/show_bug.cgi?id=1096401

I have not found recent failure of this kind in the ftp test case. Latest occurrence opensuse-Tumbleweed-DVD-x86_64-Build20180623-yast2_ncurses@64bit.

I guess it makes sense to finish #37710 first and then confirm with good statistics if this issue can still appear.

TFTP was failing due to change of workflow in yast, particularly log view was replaced by journal view. It's not clear what is going on with FTP, maybe it is related to bug reported by @dimstar.

I couldn't reproduce this issue in 100 local test runs http://dhcp128.suse.cz/tests/5035#next_previous. The only failure related to ftp test module is a mistype caused by script_run. The latest failure in o3 opensuse-Tumbleweed-DVD-x86_64-Build20180623-yast2_ncurses@64bit

Seems that again related to the SUT performance. Need to crosscheck with @okurz if we plan to run less worker instances on o3. Issue is not there for SLE.

On top, other problem for performance is that in the test suite we start a lot of services, and do not stop them, so they affect next test suites. Therefore, I would reject this ticket in favor of:
a) having same workers performance as we have for SLE
b) isolating test modules runs (e.g. with feature in os-autoinst to revert to lastgood snapshot even if test was successful if special flag is defined).

Will reject, decide for follow up ticket.

Have you set a comment on the test suite just as last time referencing a ticket about how to get rid of the timeout scale again?