[bug] Cluster can not recover after killing leader pod. #353
Description
Describe the problem
Latest version, two nodes Kill Leader Pod, Leader Pod can't recover, always CANDIDATE status, xenon logs are as follows:
2022/01/05 17:01:00.210863 api.go:285: [INFO] mysql.slave.status:&{mysql-bin.000002 8422861 false No true Yes de24d445-6e04-11ec-aa8c-4a038914a31b:1-487 de24d445-6e04-11ec-aa8c-4a038914a31b:1-487 Slave has read all relay log; waiting for more updates }
2022/01/05 17:01:00.210887 trace.go:32: [WARNING] CANDIDATE[ID:cl-222932-mysql-1.cl-222932-mysql.radondb-mysql-e2e:8801, V:30, E:1].send.requestvote.to.peer[cl-222932-mysql-0.cl-222932-mysql.radondb-mysql-e2e:8801].request.gtid[{mysql-bin.000002 8422861 false No true Yes de24d445-6e04-11ec-aa8c-4a038914a31b:1-487 de24d445-6e04-11ec-aa8c-4a038914a31b:1-487 Slave has read all relay log; waiting for more updates }]
2022/01/05 17:01:00.214574 trace.go:32: [WARNING] CANDIDATE[ID:cl-222932-mysql-1.cl-222932-mysql.radondb-mysql-e2e:8801, V:30, E:1].send.requestvote.done.to[cl-222932-mysql-0.cl-222932-mysql.radondb-mysql-e2e:8801]
2022/01/05 17:01:00.214619 trace.go:32: [WARNING] CANDIDATE[ID:cl-222932-mysql-1.cl-222932-mysql.radondb-mysql-e2e:8801, V:30, E:1].get.vote.response.from[N:cl-222932-mysql-0.cl-222932-mysql.radondb-mysql-e2e:8801, R:FOLLOWER].rsp.gtid[{ 0 false false }].retcode[ErrorInvalidRequest]
2022/01/05 17:01:00.214631 trace.go:32: [WARNING] CANDIDATE[ID:cl-222932-mysql-1.cl-222932-mysql.radondb-mysql-e2e:8801, V:30, E:1].get.vote.response.from[N:cl-222932-mysql-0.cl-222932-mysql.radondb-mysql-e2e:8801, V:0].error[ErrorInvalidRequest].but.not.downgrade.to.follower
To Reproduce
Expected behavior
Environment:
- RadonDB MySQL version: latest