Please answer these questions before submitting your issue. Thanks!
1. Minimal reproduce step (Required)
1. deploy a cluster with 3 tidb + 3 tikv, each tidb with 8cpu and 16G memory
2. import 1k warehouse data with go-tpc for ch workload
3. run the workloadgo-tpc ch --host xxx -P4000 --warehouses 1000 run -D tpcc -T 30 -t 1 --ignore-error --refresh-conn-wait 10s --time 168h0m0s
I think the implementation of rate limit doesn't reflect its design goal. I'm putting this as an enhancement as part of our internal OOM elimination work. Changing labels.
Bug Report
Please answer these questions before submitting your issue. Thanks!
1. Minimal reproduce step (Required)
2. What did you expect to see? (Required)
no tidb oom
3. What did you see instead (Required)
tidb oom
4. What is your TiDB version? (Required)
lastest master