Current logic found failed on one of prow CI run where node indeed had a core dump present but dump was not collected by "oc adm must-gather -- gather_core_dumps" command underneath a prow chain. 

It seems we may have missed core dumps collection since long time due to this race issue.

Failed logs on prod builds hereSaw a race condition where dump was tried to be copied but debug pod was already removed.

4.21

Always

    1.
    2.
    3.