client keeps using an already closed producer

#### Expected behavior
when a producer is closed on the broker side, the client should recreate a producer.
#### Actual behavior

getting message send timeout: TimeoutError on the producer side
and on the broker side I see Producer had already been closed

#### Steps to reproduce

not sure
#### System configuration
**Pulsar version**: 2.7.0
golang pulsar client version: 0.6.0

The issue names no files, tests, or reliable reproduction steps; start by tracing the Go client's producer lifecycle and broker-side close handling. Done means a broker-closed producer is recreated and message sends no longer time out, with a reproducible test covering the reported Pulsar 2.7.0 behavior.

Written by the indexing model from the issue text.