pdms rolling restart failed when enable ClusterDomain

## Bug Report

### What did you do?
1、enable ClusterDomain
2、rolling restart pdms

### What did you expect to see?
pdms rolling restart can succeed

### What did you see instead?
pdms rolling restart failed when enable ClusterDomain
`{"log":"E0703 02:34:30.994517 1 pd_ms_upgrader.go:158] TidbCluster: [endless-ha-test-pd-ms-tps-7768962-1-688/tc]' pdms upgrader: failed to transfer pdms primary to: tc-tso-1, Error response 500 URL http://tc-tso.endless-ha-test-pd-ms-tps-7768962-1-688.svc.cluster.local:2379/tso/api/v1/primary/transfer,body response: \"no valid secondary to transfer primary, from tc-tso-0.tc-tso-peer.endless-ha-test-pd-ms-tps-7768962-1-688.svc.cluster.local to tc-tso-1\"\n","container":"tidb-operator","stream":"stderr","pod":"tidb-controller-manager-test-infra-only-c9df55895-b6vhk","time":"2025-07-03T02:34:30.994569229Z","namespace":"tidb-admin"}`

### What version of PD are you using (`pd-server -V`)?
./pd-server -V
Release Version: v9.0.0-beta.2.pre-6-gb85493a
Edition: Community
Kernel Type: Classic
Git Commit Hash: b85493a5be02fff46f6325a35b09dcc98d696c46
Git Branch: HEAD
UTC Build Time: 2025-07-03 02:25:16
2025-07-03T17:14:31.612+0800 INFO k8s/client.go:140 it should be noted that a long-running command will not be interrupted even the use case has ended. For more information, please refer to https://github.com/pingcap/test-infra/discussions/129
./tidb-server -V

operator: 1.6.2-pre