TiDB日志中大量"switch region leader to specific leader due to kv return NotLeader"信息

为提高效率,提问时请提供以下信息,问题描述清晰可优先响应。

  • 【TiDB 版本】:4.0.8 w/ DM 2.0.0
  • 【问题描述】:

当前我们正在利用DM从测试库中向TiDB同步50张表的数据,实际QPS很低(大约20~40之间)。在DM Task写入的那个TiDB实例日志上发现大量的switch region leader to specific leader due to kv return NotLeader信息,且有周期性,每隔30s就会产生一批。如下:

[2020/11/25 20:01:53.508 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22624] [currIdx=2] [leaderStoreID=2]
[2020/11/25 20:01:53.508 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22624] [currIdx=2] [leaderStoreID=2]
[2020/11/25 20:01:53.508 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22624] [currIdx=2] [leaderStoreID=2]
[2020/11/25 20:01:53.508 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22593] [currIdx=0] [leaderStoreID=8]
[2020/11/25 20:01:53.508 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22593] [currIdx=0] [leaderStoreID=8]
[2020/11/25 20:01:53.508 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22593] [currIdx=0] [leaderStoreID=8]
[2020/11/25 20:01:53.510 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22552] [currIdx=0] [leaderStoreID=28]
[2020/11/25 20:01:53.510 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22552] [currIdx=0] [leaderStoreID=28]
[2020/11/25 20:01:53.510 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22552] [currIdx=0] [leaderStoreID=28]
[2020/11/25 20:01:53.510 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22552] [currIdx=0] [leaderStoreID=28]
[2020/11/25 20:01:53.779 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=23436] [currIdx=0] [leaderStoreID=14]
[2020/11/25 20:01:53.779 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=23436] [currIdx=0] [leaderStoreID=14]
[2020/11/25 20:01:53.779 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=23436] [currIdx=0] [leaderStoreID=14]
[2020/11/25 20:01:53.779 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=23436] [currIdx=0] [leaderStoreID=14]
[2020/11/25 20:01:53.779 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=23436] [currIdx=0] [leaderStoreID=14]
[2020/11/25 20:01:53.779 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=23436] [currIdx=0] [leaderStoreID=14]
[2020/11/25 20:02:23.511 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22624] [currIdx=0] [leaderStoreID=25]
[2020/11/25 20:02:23.511 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22624] [currIdx=0] [leaderStoreID=25]
[2020/11/25 20:02:23.511 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22624] [currIdx=0] [leaderStoreID=25]
[2020/11/25 20:02:23.512 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22593] [currIdx=2] [leaderStoreID=27]
[2020/11/25 20:02:23.512 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22593] [currIdx=2] [leaderStoreID=27]
[2020/11/25 20:02:23.512 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22593] [currIdx=2] [leaderStoreID=27]
[2020/11/25 20:02:23.514 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22552] [currIdx=1] [leaderStoreID=15]
[2020/11/25 20:02:23.514 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22552] [currIdx=1] [leaderStoreID=15]
[2020/11/25 20:02:23.514 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22552] [currIdx=1] [leaderStoreID=15]
[2020/11/25 20:02:23.514 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22552] [currIdx=1] [leaderStoreID=15]
[2020/11/25 20:02:53.483 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22624] [currIdx=1] [leaderStoreID=2]
[2020/11/25 20:02:53.483 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22624] [currIdx=1] [leaderStoreID=2]
[2020/11/25 20:02:53.484 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22624] [currIdx=1] [leaderStoreID=2]
[2020/11/25 20:02:53.484 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22593] [currIdx=0] [leaderStoreID=8]
[2020/11/25 20:02:53.484 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22593] [currIdx=0] [leaderStoreID=8]
[2020/11/25 20:02:53.484 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22593] [currIdx=0] [leaderStoreID=8]
[2020/11/25 20:02:53.486 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22552] [currIdx=0] [leaderStoreID=28]
[2020/11/25 20:02:53.486 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22552] [currIdx=0] [leaderStoreID=28]
[2020/11/25 20:02:53.486 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22552] [currIdx=0] [leaderStoreID=28]
[2020/11/25 20:02:53.486 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22552] [currIdx=0] [leaderStoreID=28]
[2020/11/25 20:02:53.624 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=23436] [currIdx=2] [leaderStoreID=12]
[2020/11/25 20:02:53.624 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=23436] [currIdx=2] [leaderStoreID=12]
[2020/11/25 20:02:53.624 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=23436] [currIdx=2] [leaderStoreID=12]
[2020/11/25 20:02:53.624 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=23436] [currIdx=2] [leaderStoreID=12]
[2020/11/25 20:02:53.624 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=23436] [currIdx=2] [leaderStoreID=12]
[2020/11/25 20:03:07.257 +08:00] [INFO] [client_batch.go:655] ["recycle idle connection"] [target=10.0.46.173:20162]
[2020/11/25 20:03:07.257 +08:00] [INFO] [client_batch.go:655] ["recycle idle connection"] [target=10.0.46.178:20162]
[2020/11/25 20:03:23.501 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22624] [currIdx=0] [leaderStoreID=25]
[2020/11/25 20:03:23.501 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22624] [currIdx=0] [leaderStoreID=25]
[2020/11/25 20:03:23.501 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22624] [currIdx=0] [leaderStoreID=25]
[2020/11/25 20:03:23.502 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22593] [currIdx=2] [leaderStoreID=27]
[2020/11/25 20:03:23.502 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22593] [currIdx=2] [leaderStoreID=27]
[2020/11/25 20:03:23.502 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22593] [currIdx=2] [leaderStoreID=27]
[2020/11/25 20:03:23.504 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22552] [currIdx=1] [leaderStoreID=15]
[2020/11/25 20:03:23.504 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22552] [currIdx=1] [leaderStoreID=15]
[2020/11/25 20:03:23.504 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22552] [currIdx=1] [leaderStoreID=15]
[2020/11/25 20:03:23.504 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22552] [currIdx=1] [leaderStoreID=15]
[2020/11/25 20:03:23.612 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=23436] [currIdx=0] [leaderStoreID=23]
[2020/11/25 20:03:23.612 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=23436] [currIdx=0] [leaderStoreID=23]
[2020/11/25 20:03:53.486 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22624] [currIdx=1] [leaderStoreID=2]
[2020/11/25 20:03:53.486 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22624] [currIdx=1] [leaderStoreID=2]
[2020/11/25 20:03:53.486 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22624] [currIdx=1] [leaderStoreID=2]
[2020/11/25 20:03:53.486 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22593] [currIdx=0] [leaderStoreID=30]
[2020/11/25 20:03:53.487 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22593] [currIdx=0] [leaderStoreID=30]
[2020/11/25 20:03:53.487 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22593] [currIdx=0] [leaderStoreID=30]
[2020/11/25 20:03:53.487 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22593] [currIdx=0] [leaderStoreID=30]
[2020/11/25 20:03:53.487 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22593] [currIdx=0] [leaderStoreID=30]
[2020/11/25 20:03:53.488 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22552] [currIdx=0] [leaderStoreID=28]
[2020/11/25 20:03:53.488 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22552] [currIdx=0] [leaderStoreID=28]
[2020/11/25 20:03:53.488 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22552] [currIdx=0] [leaderStoreID=28]
[2020/11/25 20:03:53.488 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=22552] [currIdx=0] [leaderStoreID=28]
[2020/11/25 20:03:53.630 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=23436] [currIdx=1] [leaderStoreID=14]
[2020/11/25 20:03:53.630 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=23436] [currIdx=1] [leaderStoreID=14]
[2020/11/25 20:03:53.630 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=23436] [currIdx=1] [leaderStoreID=14]
[2020/11/25 20:03:53.630 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=23436] [currIdx=1] [leaderStoreID=14]
[2020/11/25 20:03:53.630 +08:00] [INFO] [region_cache.go:839] ["switch region leader to specific leader due to kv return NotLeader"] [regionID=23436] [currIdx=1] [leaderStoreID=14]

停掉DM Task之后,此信息就不再输出。集群中有3个TiDB实例和18(6 * 3)个TiKV实例。

请问是什么原因?感谢~

1、少量的 该错误并不影响使用,主要意思是:发送的读写请求到指定的 region 后,发现该 region 已经不是 leader,数据库会自动重试,所以并不会影响使用,但如果数量较多,可能会影响执行效率
2、目前怀疑 集群有较多的 调度,等集群稳定后,情况会变好,这个属于正常调度,暂时无法确定是否异常

好的,目前确实没有影响到读写,我们再观察一段时间,感谢~

:ok_hand:有问题可以随时开帖

此话题已在最后回复的 1 分钟后被自动关闭。不再允许新回复。