Stale ETCD client endpoints remain in healthyChecker
#7226
Labels
affects-7.5
This bug affects the 7.5.x(LTS) versions.
severity/major
type/bug
The issue is confirmed as a bug.
Bug Report
What did you do?
I tried to upgrade a 3-node PD cluster's traffic from plaintext HTTP to HTTPS, at some stage, there will be 6 client endpoints, like:
When I bring those HTTP endpoints offline, the health checker still blames the HTTP endpoints are not being healthy, which might stop checking those endpoints.
What did you expect to see?
Check the health status for those endpoints reported from the ETCD cluster.
What did you see instead?
Stale endpoints are checked forever.
What version of PD are you using (
pd-server -V
)?master
The text was updated successfully, but these errors were encountered: