从Nacos 1.3.2升级到Nacos 2.0.1，持续打印upgrade check resul

3个节点从Nacos 1.3.2升级到Nacos 2.0.1，按照升级文档操作后，3个节点都持续打印upgrade check result false。

定位到完成Check工作的类是com.alibaba.nacos.naming.core.v2.upgrade.UpgradeJudgement类，改造了checkServiceAndInstanceNumber方法打印Check工作的现场信息。

改造后checkServiceAndInstanceNumber方法如下：

private boolean checkServiceAndInstanceNumber() { boolean result = serviceManager.getServiceCount() == MetricsMonitor.getDomCountMonitor().get(); result &= serviceManager.getInstanceCount() == MetricsMonitor.getIpCountMonitor().get();

    Loggers.SRV_LOG.error("TAG: ServiceCount-{}-{}, InstanceCount-{}-{}", serviceManager.getServiceCount(),
        MetricsMonitor.getDomCountMonitor().get(), serviceManager.getInstanceCount(),
        MetricsMonitor.getIpCountMonitor().get());

    return result;
}

基于改造后的JAR包运行后，发现日志如下：

2021-05-10 14:33:50,785 ERROR TAG: ServiceCount-0-0, InstanceCount-0-0 2021-05-10 14:33:55,775 ERROR TAG: ServiceCount-38-0, InstanceCount-251-251 2021-05-10 14:34:00,774 ERROR TAG: ServiceCount-38-38, InstanceCount-251-251 2021-05-10 14:34:05,774 ERROR TAG: ServiceCount-38-38, InstanceCount-251-251 2021-05-10 14:34:10,774 ERROR TAG: ServiceCount-38-38, InstanceCount-251-251 2021-05-10 14:34:15,774 ERROR TAG: ServiceCount-38-38, InstanceCount-251-251 2021-05-10 14:34:20,776 ERROR TAG: ServiceCount-38-38, InstanceCount-251-251 2021-05-10 14:34:25,774 ERROR TAG: ServiceCount-38-38, InstanceCount-251-219 2021-05-10 14:34:30,776 ERROR TAG: ServiceCount-38-38, InstanceCount-251-182 2021-05-10 14:34:35,774 ERROR TAG: ServiceCount-38-38, InstanceCount-251-182 2021-05-10 14:34:40,774 ERROR TAG: ServiceCount-38-38, InstanceCount-251-182 2021-05-10 14:34:45,774 ERROR TAG: ServiceCount-38-38, InstanceCount-251-182 2021-05-10 14:34:50,774 ERROR TAG: ServiceCount-38-38, InstanceCount-251-182 2021-05-10 14:34:55,774 ERROR TAG: ServiceCount-38-38, InstanceCount-251-214 2021-05-10 14:35:00,774 ERROR TAG: ServiceCount-38-38, InstanceCount-251-214 2021-05-10 14:35:05,774 ERROR TAG: ServiceCount-38-38, InstanceCount-251-214

也就是说通不过Check的原因是serviceManager.getInstanceCount() == MetricsMonitor.getIpCountMonitor().get()持续返回false.

问题1：请问如何解决以上问题？

问题2：另外，还有1个困惑：假如当前升级成功+关闭双写后，下次重启一个节点是否仍然会做以上检查，如果彼时检查通不过，会不会产生数据一致性问题，而不能正常工作？

原提问者GitHub用户dslztx