The port this QNode will run on
The host this QNode will run on
Number of Thrift connections allocated as a connection pool in each QNode.
Whether this QNode should find the next available port in case "dnode.port" is busy or fail otherwise.
The number of succeessfully deployed versions that will be kept in the system (per tablespace)
The timeout for a global deploy in seconds. -1 for disabling timeout
The number of seconds to wait before checking each time if a DNode has failed or if timeout has ocurred in the
middle of a deploy
A fixed amount of time (seconds) that this QNode will wait before taking certain actions. For example, a QNode may decide to
re-balance under-replicated partitions. But if we are at cluster start, some DNodes may still be connecting or are
to connect in the near future. For that it is useful to wait this "warming time" before deciding such things.
The TTL (in seconds) of balance actions: if they take more than this number of seconds they will be evicted.
This is necessary because there are a number of race conditions that might block certain actions (i.e. final DNode
going down in the middle of the transfer).
Whether or not to enable automatic replica balancing.
If automatic replica balancing is enabled, the system will become highly available so that
if one replica is down, it will copy other replicas to another DNode so that the affected
partitions don't remain under-replicated.
As a downside, replica balancing might use more disk space than desired.