Do the primary and secondary subclusters need to have the same node count and the same sharding footprints?@dsprogis @ybei
Eon Mode: Past, Present, Future
Shard count is an attribute of the database, once set, it never changes. Subclusters are most efficient when the shards are evenly distributed across nodes because the query will only run as fast as the slowest node (the node with the most shard subscriptions). Thus, the number of nodes should be an factor of the shard count. A shard count of 12 is good because its factors are 1, 2, 3, 4, 6, and 12 which gives you a lot of options. Importantly, because shard count establishes the maximum number of nodes that will participate in a query, you should size it high enough for your most complex, long running queries. This is the case regardless of whether the subcluster is primary or secondary.
Can't find what you're looking for? Search the Vertica Documentation, Knowledge Base, or Blog for more information.