Note: for the most recent updates on the topic of cluster membership and node lifecycle, please also check:

Also, see related:

Joining a Cluster: discovery URL, and more

First, some basic facts. AIStore clusters can be deployed with an arbitrary number of AIStore proxies. Each proxy/gateway implements RESTful API and provides full access to objects stored in the cluster. Each proxy collaborates with all other proxies to perform majority-voted HA failovers (section Highly Available Control Plane.

All electable proxies are functionally equivalent. The one that is elected as primary is, among other things, responsible to join nodes to the running cluster.

To facilitate node-joining in presence of disruptive events, such as:

  • network failures, and/or

  • partial or complete loss of local copies of aistore metadata (e.g., cluster maps)

  • to still be able to reconnect and restore operation, we also provide so called original and discovery URLs in the cluster configuration.

The latter is versioned, replicated, protected and distributed - solely by the elected primary.

March 2024 UPDATE: starting v3.23, the original URL does not track the “original” primary. Instead, the current (or currently elected) primary takes full responsibility for updating both URLs with the single and singular purpose: optimizing time to join or rejoin cluster.

For instance:

When and if an HA event triggers automated failover, the role of the primary will be automatically assumed by a different proxy/gateway, with the corresponding cluster map (Smap) update getting synchronized across all running nodes.

A new node, however, could potentially experience a problem when trying to join an already deployed and running cluster - simply because its configuration may still be referring to the old primary. The original and discovery URLs (see AIStore configuration) are precisely intended to address this scenario.