The only considered aspect, when configuring Syncope for high-availability, is related to OpenJPA's remote commit provider, at least according to the current documentation.
However, another component is relevant within this regard, e.g. the Quartz scheduler, which is currently simply setup by default for clustering.
With such configuration, all cluster nodes are equally selectable for processing jobs; it would be nice, though, to gain some control over this aspect, e.g. to be able to restrict the nodes where jobs can run.
For example, one can have 3 Syncope core nodes configured with OpenJPA remote commit provider, and set 2 of them for processing REST requests, leaving the third dedicated for running jobs (e.g. Pull).