When autoid service panic unexpected, the recovering time is a bit too long #46454
Labels
affects-6.5
This bug affects the 6.5.x(LTS) versions.
affects-7.1
This bug affects the 7.1.x(LTS) versions.
sig/sql-infra
SIG: SQL Infra
type/enhancement
The issue or PR belongs to an enhancement.
Enhancement
When autoid service leader panic, the follower need to wait the lease timeout to take over the leader's role.
The lease is 60s, which is a bit too long...
For DDL service, it's also 60s, and it's the default value https://pkg.go.dev/go.etcd.io/etcd/clientv3/concurrency#WithTTL
DDL is offline service, so 60s is tolerable.
But autoid allocating is online service, the custome may not expect it take so long to recover.
The text was updated successfully, but these errors were encountered: