PD端口失去响应

tidb版本3.0

登陆PD节点,发现如下信息:
[2020/09/23 01:07:12.420 +08:00] [WARN] [v3_server.go:700] [“timed out waiting for read index response (local node might have slow network)”] [timeout=11s]
[2020/09/23 01:07:08.165 +08:00] [WARN] [util.go:144] [“apply request took too long”] [took=2.32710494s] [expected-duration=100ms] [prefix=“read-only range "
] [request=“key:”/tidb-binlog/v1/drainers/db-tidb-1:8249” "] [response=] [error=“context deadline exceeded”]
[2020/09/23 01:07:12.610 +08:00] [WARN] [util.go:144] [“apply request took too long”] [took=2.738279511s] [expected-duration=100ms] [prefix="read-only range
“] [request=“key:”/tidb/store/gcworker/saved_safe_point” “] [response=] [error=“context deadline exceeded”]
[2020/09/23 01:07:31.013 +08:00] [WARN] [v3_server.go:657] [“failed to get read index from Raft”] [error=“context deadline exceeded”]
[2020/09/23 01:07:31.072 +08:00] [WARN] [probing_status.go:70] [“prober detected unhealthy status”] [round-tripper-name=ROUND_TRIPPER_SNAPSHOT] [remote-peer-
id=49ad4e2050f96f3a] [rtt=1.94094326s] [error=“dial tcp 10.33.66.62:2380: i/o timeout”]
[2020/09/23 01:07:32.202 +08:00] [WARN] [probing_status.go:70] [“prober detected unhealthy status”] [round-tripper-name=ROUND_TRIPPER_RAFT_MESSAGE] [remote-p
eer-id=49ad4e2050f96f3a] [rtt=1.256516436s] [error=“dial tcp 10.33.66.62:2380: i/o timeout”]
[2020/09/23 01:07:34.939 +08:00] [WARN] [util.go:144] [“apply request took too long”] [took=3.674655056s] [expected-duration=100ms] [prefix=“read-only range
“] [request=“key:”/tidb-binlog/v1/pumps/” range_end:”/tidb-binlog/v1/pumps0” "] [response=] [error=“context deadline exceeded”]
[2020/09/23 01:07:37.652 +08:00] [WARN] [probing_status.go:86] [“prober found high clock drift”] [round-tripper-name=ROUND_TRIPPER_SNAPSHOT] [remote-peer-id=
49ad4e2050f96f3a] [clock-drift=1.94094326s] [rtt=3.427876755s] [error=“dial tcp 10.33.66.62:2380: i/o timeout”]
[2020/09/23 01:07:38.561 +08:00] [ERROR] [redirector.go:114] [“request failed”] [error=“Get http://10.33.66.62:2379/pd/ping: context canceled”]

时间点刚好是mydumper备份期间。由于前天新增了一张20G的大表。23,24号凌晨备份期间均出现该
错误信息。PD节点通信中断。备份整体时间为25分钟。表中有几个varchar(1000)的字段。字段过大会导致PD节点出问题么?

1、当前 TiDB 部署的拓扑结构辛苦提供下,尤其是混部的情况,如 TiDB,TiKV 以及 PD 间出现混部~

2、在备份时,发起 mydumper 备份命令的服务器的网卡流量以及网卡配置(是千兆还是万兆网卡)~