Ticket 49463 - After cleanALLruv, there is a flow of keep alive DEL
When cleanAllRuv is launched, it spawn cleanAllRuv on all replicas.
Each replica will clean its changelog and database RUV AND in addition
will DEL the keep alive entry of the target ReplicaID.
So for the same entry (keep alive) there will be as many DEL as there are replicas
This flow of DEL is useless as only one DEL is enough.
In addition because of https://pagure.io/389-ds-base/issue/49466, replication may
loop on each of those DELs.
The fix is only to prevent the flow of DEL.
It adds a flag ('original_task') in the task payload.
The server receiving the task (replica_execute_cleanall_ruv_task) flags the
task as 'original_task'.
In the opposite, the propagated cleanAllRuv (multimaster_extop_cleanruv) does
not flag the task as 'original_task'
Only original task does the DEL of the keep alive entry.
Note the propageted payload (extop) is not changed. In a mixed version
environment "old" servers will DEL the keep alive and flow can still happen
Reviewed by: Ludwig Krispenz
Platforms tested: F23
Flag Day: no
Doc impact: no