[ClusterLabs] Q: crm_resource -C

[Ulrich Windl]

Hi!

It seems there is a change in semantics (or a bug?) from SLES11 to SLES12/SLES15 regarding crm_resource -C -r ...:
I had a cloned resource that failed startup due to a missing configuration file:

Failed Resource Actions:
  * prm_test_raid_monitor_0 on h19 'not installed' (5): call=17, status='complete', exitreason='Configuration file [/etc/mdadm/mdadm.conf] does not exist, or can not be opened!', last-rc-change='2020-11-26 15:40:30 +01:00', queued=0ms, exec=26ms
  * prm_test_raid_monitor_0 on h18 'not installed' (5): call=17, status='complete', exitreason='Configuration file [/etc/mdadm/mdadm.conf] does not exist, or can not be opened!', last-rc-change='2020-11-26 15:40:30 +01:00', queued=0ms, exec=29ms

I cleaned up the resource sucessfully in SLES15 SP2, but the error condition did not reset, and no new start attempt happened:
# crm_resource -r prm_test_raid -C -n h18
Cleaned up prm_test_raid:0 on h16
Cleaned up prm_test_raid:1 on h19
Cleaned up prm_test_raid:1 on h18
Cleaned up prm_test_raid:2 on h16
Cleaned up prm_test_raid:2 on h19
Cleaned up prm_test_raid:2 on h18

Intended, or a bug? The manual description for "-C" says: "If resource has any past failures, clear its history and fail count."

Regards,
Ulrich