nacos 2.0.3集群升级jvm到4G，重启导致集群不可用，再次扩容到8G后才恢复，求助原因

[Roronoa-Zoro]

3节点的集群，前面挂了一个LB，先把一台从LB摘流，然后jvm从2G改到4G重启，整个集群变得不可用，节点的状态持续变化UP/Down，然后日志里面报了很多下面的错误，最后节点扩容后，jvm改成8G，重启才恢复。
恢复后看单机IP连接数1万，服务梳理接近3000，
有大佬帮忙分析下原因吗，就是单纯的内存不足吗？摘流后，应该新节点不会被连接吧

日志的话，报了很多这种，
Server check fail
Client not connected
DistroDataRequest send request fail
send request fail
old raft protocol already stop

2026-08-05 16:13:07,500 INFO [Cluster-10.3.10.32:8848] success to connect a server [10.3.10.32:8848],connectionId=1785917587384_10.3.10.31_47364

2026-08-05 16:13:07,500 INFO [Cluster-10.3.10.32:8848] Abandon prev connection ,server is 10.3.10.32:8848, connectionId is 1777201307602_10.3.10.31_52
856

2026-08-05 16:13:07,549 ERROR [Cluster-10.3.10.32:8848]Send request fail, request=DistroDataRequest{headers={}, requestId='null'}, retryTimes=2,errorMe
ssage=Client not connected.

2026-08-05 16:13:07,552 ERROR [Cluster-10.3.10.32:8848]Send request fail, request=DistroDataRequest{headers={}, requestId='null'}, retryTimes=0,errorMe
ssage=Client not connected.

2026-08-05 16:13:07,553 INFO [Cluster-10.3.10.32:8848] Server check success,currentServer is10.3.10.32:8848

2026-08-05 16:13:07,600 ERROR [Cluster-10.3.10.32:8848]Send request fail, request=DistroDataRequest{headers={}, requestId='null'}, retryTimes=2,errorMe
ssage=Client not connected.

2026-08-05 16:13:07,600 ERROR [Cluster-10.3.10.32:8848]Send request fail, request=DistroDataRequest{headers={}, requestId='null'}, retryTimes=0,errorMe
ssage=Client not connected.

2026-08-05 16:13:07,984 ERROR [Cluster-10.3.10.32:8848]Send request fail, request=DistroDataRequest{headers={}, requestId='null'}, retryTimes=0,errorMe
ssage=Client not connected.

看线程，很多都是这种状态
http-nio-8848-exec-*
-> com.alibaba.nacos.naming.web.DistroFilter.doFilter
-> com.alibaba.nacos.common.http.client.NacosRestTemplate.exchange
-> Apache HttpClient
-> AbstractConnPool.getPoolEntryBlocking