pd client panics with "[tso] timestamp fallback"

[overvenus]

Bug Report

[2025/02/12 11:20:05.393 +00:00] [PANIC] [dispatcher.go:528] ["[tso] timestamp fallback"] [keyspace=4294967295] [last-ts="(1739359206335, 2)"] [cur-ts="(1739359205352, 4)"] [last-tso-server=node-4.node-peer.jepsen-tps-7781314-1-684:2379] [cur-tso-server=node-0.node-peer.jepsen-tps-7781314-1-684:2379] [last-keyspace-group-in-request=0] [cur-keyspace-group-in-request=0] [last-keyspace-group-in-response=0] [cur-keyspace-group-in-response=0] [last-response-received-at=2025/02/12 11:20:05.357 +00:00] [cur-response-received-at=2025/02/12 11:20:05.393 +00:00] [last-stream-id=node-4.node-peer.jepsen-tps-7781314-1-684:2379-191] [cur-stream-id=node-0.node-peer.jepsen-tps-7781314-1-684:2379-192]

Test: https://tcms.pingcap.net/dashboard/executions/plan/7781314

The panic was found on a cluster without the recent fix #9056 , maybe they are related.

What did you do?

Run jepsen tests.

What did you expect to see?

No panic.

What version of PD are you using (pd-server -V)?

TiDB: pingcap/tidb@c292ec6
PD: debceaf

[rleungx]

The reason is the same as #9051