Fix procedure of masterNode determine in MySQL upgrade pipeline

If stopped servers have the same seqno in /var/lib/mysql/grastate.dat
then master became the last checked instead of the last stopped

Current behavior changed to use the getGaleraLastShutdownNode method

PROD-36755

Change-Id: I1dfbb51f2c4c1b3383c329db8facfc6789332e09
(cherry picked from commit 34603b305344258cf618f9e83b591024ac38d063)
diff --git a/openstack-galera-upgrade.groovy b/openstack-galera-upgrade.groovy
index f124051..64dbbad 100644
--- a/openstack-galera-upgrade.groovy
+++ b/openstack-galera-upgrade.groovy
@@ -133,17 +133,7 @@
       }
     }
 
-    for (target in upgradeTargets) {
-         out = salt.cmdRun(env, target,  'cat /var/lib/mysql/grastate.dat | grep "seqno" | cut -d ":" -f2', true, null, false).get('return')[0].values()[0].replaceAll('Salt command execution success', '').trim()
-         common.infoMsg("Get seqno: ${out} for node ${target}")
-         if (!out.isNumber()){
-             out = -2
-         }
-        targetSecMapping[out.toInteger()] = target
-        secNoList.add(out.toInteger())
-    }
-
-    def masterNode = targetSecMapping[secNoList.max()]
+    def masterNode = salt.getMinionsSorted(env, galera.getGaleraLastShutdownNode(env))[0]
     common.infoMsg("Master node is: ${masterNode}")
 
     // Make sure we start upgrade always from master node