Min Zhou

The external IP of tidb service load balancer was stuck on ``. ``` Warning SyncLoadBalancerFailed 5m18s (x7 over 10m) service-controller Error syncing load balancer: failed to ensure load balancer: Retriable:...

I'm not sure if this patch is intended to solve the query latency spike issue caused by rolling restart of PD. If so, this patch may not solve the problem....

> @coderplay Is it only a P100/long tail latency, or all TiDB queries that are affected in your experiment? This is P100/long tail latency, no obvious spikes on P95. Theoretically,...

After looking at the PD code, the `pd-ctl transfer leader` and the `resign` in this PR are actually equivalent. `Resign` simply allows PD to randomly select another PD as the...

Thanks a lot Haoyi, for reviewing this feature! Looking forward to the scala implementation.

@DanielZhangQD The way I understand it as `managed` is that the PV is provisioned and maintained by the tidb-operator. TiDB-operator adds many labels and annotations to the PV it manages....

My overall impression of this PR is that we are leveraging certain undocumented invariants, e.g. sole append-only writer plus one reader instance per thread, in the upper layers to ensure...

In a scenario where a fast disk wrote transactions 1, 2, and 3, while the slow disk only recorded transaction 1. If the fast disk is lost somehow, and only...