[Master][Bug] concurrent map r/w in roundRobinQueueSelector causes panic

[czah]

1. Please describe the issue you observed:

   • What did you do (The steps to reproduce)?
     Run a producer and start producing messages.

   • What did you expect to see?
     Everything is working well.

   • What did you see instead?
     It panicked with a error message (as shown below):

fatal error: concurrent map read and map write

goroutine 4877278 [running]:
runtime.throw(0xe50ffa, 0x21)
	/usr/local/go/src/runtime/panic.go:774 +0x72 fp=0xc000be4b20 sp=0xc000be4af0 pc=0x431182
runtime.mapaccess2_faststr(0xd0cea0, 0xc00031c810, 0xc0004b2f14, 0x5, 0x100000000000000, 0xd6cf00)
	/usr/local/go/src/runtime/map_faststr.go:116 +0x48f fp=0xc000be4b90 sp=0xc000be4b20 pc=0x414c5f
github.com/apache/rocketmq-client-go/v2/producer.(*roundRobinQueueSelector).Select(0xc0002b5700, 0xc0002663f0, 0xc001204400, 0x6, 0x8, 0x101)
	/go/pkg/mod/github.com/apache/rocketmq-client-go/[email protected]/producer/selector.go:77 +0x5f fp=0xc000be4be8 sp=0xc000be4b90 pc=0xc2ae5f
github.com/apache/rocketmq-client-go/v2/producer.(*defaultProducer).selectMessageQueue(0xc000234ea0, 0xc0002663f0, 0xc00146c1e0)
	/go/pkg/mod/github.com/apache/rocketmq-client-go/[email protected]/producer/producer.go:396 +0x182 fp=0xc000be4c90 sp=0xc000be4be8 pc=0xc2a952
...

It was caused by an unprotected read op on the map named indexer in roundRobinQueueSelector.

2. Please tell us about your environment:

   • What is your OS?
     A Docker container based on the CentOS 7.3 image (host kernel version: 4.9.2-3.0.0.std7b.el7.5.x86_64)

   • What is your client version?
     v2.1.0-rc6

   • What is your RocketMQ version?
     v4.8

3. Other information (e.g. detailed explanation, logs, related issues, suggestions on how to fix, etc):