action #17442
Updated by okurz about 7 years ago
## observation tests like zypper_lifecycle, sshd, ssh_sleanup, sshxterm and other started to fail much more than some time ago. It is not new issue, but this failures were not so frequent as now. All fails are caused by typing to soon or letters get lost because of very low performance. #16738 is somehow related. ## reproducible seems to happen in some jobs ## problem * H1 REJECTED worker overbooking * H2 qemu loosing keys * H2.1 REJECTED loosing keys in keyboard buffer -> rejected: we would expect "key event buffer full" messages and an early fail in the jobs * H2.2 loosing keys * H3 performance regression in the product * H4 REJECTED network -> only affecting x11 tests and not console tests so why should we loose just VNC typing events there? (okurz) * H5 REJECTED package updates on the workers -> no updates on worker since 3 weeks (rbrown) ## suggestion * E1-1 check workers of last examples -> O1-1 linked jobs are spread over openqaworker2, …3, …5, …6 -> REJECT H1 * E3-1 crosscheck older build ## further details Few examples: https://openqa.suse.de/tests/793680#step/yast2_snapper/5 https://openqa.suse.de/tests/793974#step/zypper_lifecycle/55 <del>https://openqa.suse.de/tests/793716#step/ssh_cleanup/3</del> --> #14826 <del>https://openqa.suse.de/tests/793683#step/sshd/16</del> --> #14826 http://openqa.suse.de/t796211 https://openqa.suse.de/tests/796232#step/gnome_terminal/7 https://openqa.suse.de/tests/795241 (Build 0260) ## workaround retrigger job