We observed problem with intermittently failing replication jobs between federated catalyst stores with following error:
[Critical] From: CSM@hostname "replication_name" Time: 14/01/2016 10:46:50
Replication error: Changed whilst queued or paused.
Problem was identified as a bug by StoreOnce team, it arose because the nodes within the appliance did not all have the same system time set, which meant consistency checks failed due to a “last modified” timestamp varying on the individual portions of some Federated objects.
A future version of StoreOnce software will resolve this issue completely, by improving the accuracy of NTP for multimode systems and/or relaxing the consistency check for Federated Objects.
If the problem is seen case should be opened with StoreOnce support team