[jms] when using failover [re]connection attempts should stop upon non-temporary SASL failure

When using a failover URI, the client continues to try and [re]connect upon encountering SASL failure for a given server. Realistically different servers in the failover list should behave similarly here, so continuing to attempt connection is likely to be futile and lead to a delay signalling the issue when the allowed attempts run out. When the SASL process fails with an outcome other than sys-temp, further [re]connection attempts should stop.