vearch集群重启ps选主失败

[hanqiushi]

类似 #724 ，我用最新的 vearch/vearch:latest 还是能复现这个问题。

配置如下：
三副本，外置etcd，每个副本参数都是 all (master，ps，client都包括)，想做成高可用。

复现步骤如下：
建表，replica为3，插入一些数据。
把三个节点都挂掉，然后一起重启，发现选主还是不成功

下面是一台机器的异常log
2023-09-14 06:50:10,934 :0: DEBUG: raft[3] became candidate at term 674.
2023-09-14 06:50:10,934 :0: DEBUG: raft[3] became candidate at term 674.
2023-09-14 06:50:10,934 :0: DEBUG: raft[3] received vote from 1 at term 674.
2023-09-14 06:50:10,934 :0: DEBUG: raft[3] received vote from 1 at term 674.
2023-09-14 06:50:10,934 :0: DEBUG: [raft->campaign][3 logterm: 0, index: 0] sent vote request to 2 at term 674. raftFSM[0xc000ed8000]
2023-09-14 06:50:10,934 :0: DEBUG: [raft->campaign][3 logterm: 0, index: 0] sent vote request to 2 at term 674. raftFSM[0xc000ed8000]
2023-09-14 06:50:10,934 :0: DEBUG: [raft->campaign][3 logterm: 0, index: 0] sent vote request to 3 at term 674. raftFSM[0xc000ed8000]
2023-09-14 06:50:10,934 :0: DEBUG: [raft->campaign][3 logterm: 0, index: 0] sent vote request to 3 at term 674. raftFSM[0xc000ed8000]
2023-09-14 06:50:10,934 :0: DEBUG: [Transport] get connection[Replicate] to 3[] failed,error is: cannot get node network information, nodeID=[3]
2023-09-14 06:50:10,934 :0: DEBUG: [Transport] get connection[Replicate] to 3[] failed,error is: cannot get node network information, nodeID=[3]
2023-09-14 06:50:10,934 :0: DEBUG: [Transport] get connection[Replicate] to 2[] failed,error is: cannot get node network information, nodeID=[2]
2023-09-14 06:50:10,934 :0: DEBUG: [Transport] get connection[Replicate] to 2[] failed,error is: cannot get node network information, nodeID=[2]

[hanqiushi]

2023-09-14 06:50:10,934 :0: DEBUG: [Transport] get connection[Replicate] to 3[] failed,error is: cannot get node network information, nodeID=[3]

这句log跟正常的log相比，3[] 这里数据是空的， 正常的是有数据的，如 ip:8898，感觉像是config,toml数据没有load进来或者处理过程中出错了。

[zcdb]

建议使用单模块的方式验证，all的方式仅仅是简单的功能测试