You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
After 1 tikv server down when benchmark test, tidb will output tens of thousands log about region_request in a matter of minutes, follwing is an example in my test:
Too much log about throwing pseudo region error due to no replica available
Issue-1: Too much log when 1 tikv node down
After 1 tikv server down when benchmark test, tidb will output tens of thousands log about
region_request
in a matter of minutes, follwing is an example in my test:throwing pseudo region error due to no replica available
DataIsNotReady
, this should be fix by reduce unnecessary log when DataIsNotReady #1304send request failed, err: context canceled
send request on region error failed
Issue-2: log is not friendly to diagnostic slow queries
After 1 tikv server down when benchmark test, we will find some slow queries in TiDB dashboard:
Then we may use
Connection ID
andtxn timestamp
to grep tidb log, then got following log:The problem is:
throwing pseudo region error due to no replica available
.retry-times=1
, what is the rest of the rpc error.The text was updated successfully, but these errors were encountered: