[Rollup] Fix FullClusterRestart test

We need to wait for the job to fully initialize and start before
we can attempt to stop it.  If we don't, it's possible for the stop
API to be called before the persistent task is fully loaded and it'll
throw an exception.

Closes #32773

Author: polyfractal

x-pack/qa/full-cluster-restart/src/test/java/org/elasticsearch/xpack/restart/FullClusterRestartIT.java

@@ -345,7 +345,6 @@ public void testRollupAfterRestart() throws Exception {
         }
     }
 
-    @AwaitsFix(bugUrl="https://github.com/elastic/elasticsearch/issues/32773")
     public void testRollupIDSchemeAfterRestart() throws Exception {
         assumeTrue("Rollup can be tested with 6.3.0 and onwards", oldClusterVersion.onOrAfter(Version.V_6_3_0));
         assumeTrue("Rollup ID scheme changed in 6.4", oldClusterVersion.before(Version.V_6_4_0));
@@ -413,6 +412,8 @@ public void testRollupIDSchemeAfterRestart() throws Exception {
             indexRequest.setJsonEntity("{\"timestamp\":\"2018-01-02T00:00:01\",\"value\":345}");
             client().performRequest(indexRequest);
 
+            assertRollUpJob("rollup-id-test");
+
             // stop the rollup job to force a state save, which will upgrade the ID
             final Request stopRollupJobRequest = new Request("POST", "_xpack/rollup/job/rollup-id-test/_stop");
             Map<String, Object> stopRollupJobResponse = entityAsMap(client().performRequest(stopRollupJobRequest));