Andrei Dan bbc57828c4
ILM fix retry delete action test (#54809) (#54895)
Asserting on the failed_step field from the explainAPI can produce flakiness
because the ILM state is moved back and forth between the (failing) step and
the ERROR step (as the workflow is retry, fail then move to ERROR step,
move back to the (failing) step,  retry, fail, etc) and the failed_step
information is only available whilst in the ERROR state.

Unmute other tests as they were collateral failures
A read-only index could not be deleted in the wipeCluster phase and caused
these failures

(cherry picked from commit 99a6d57aeb3cf11abc38b514f38a96bb1612e357)
Signed-off-by: Andrei Dan <andrei.dan@elastic.co>
2020-04-07 15:55:56 +01:00
..