Improve TiKV wait strategy for v6 and v7 variants

- Replace 'TiKV started' log wait with more reliable indicators:
  - Wait for 'succeed to update max timestamp' log (appears 3+ times when TiKV is active)
  - Wait for status port 20180 to be listening
- Increase timeout from 120s to 180s for better reliability
- This addresses timeout issues where TiKV was running but wait condition wasn't matching

Author: zph

mysql/testcontainers_helper.go

@@ -428,8 +428,12 @@ func startTiDBCluster(ctx context.Context, t *testing.T, version string) *TiDBTe
 					},
 				}
 			},
-			WaitingFor: wait.ForLog("TiKV started").
-				WithStartupTimeout(120 * time.Second),
+			WaitingFor: wait.ForAll(
+				// Wait for TiKV to connect to PD and start serving
+				wait.ForLog("succeed to update max timestamp").
+					WithOccurrence(3), // Wait for at least 3 region updates
+				wait.ForListeningPort("20180/tcp"), // Status port
+			).WithStartupTimeout(180 * time.Second),
 		},
 		Started: true,
 	})
@@ -554,8 +558,12 @@ func startSharedTiDBCluster(version string) (*TiDBTestCluster, error) {
 					},
 				}
 			},
-			WaitingFor: wait.ForLog("TiKV started").
-				WithStartupTimeout(120 * time.Second),
+			WaitingFor: wait.ForAll(
+				// Wait for TiKV to connect to PD and start serving
+				wait.ForLog("succeed to update max timestamp").
+					WithOccurrence(3), // Wait for at least 3 region updates
+				wait.ForListeningPort("20180/tcp"), // Status port
+			).WithStartupTimeout(180 * time.Second),
 		},
 		Started: true,
 	})