snapshot: don't schedule next snapshot job for a removed volume (apache#8735)

* snapshot: don't schedule next snapshot job for a removed volume

When management server starts, it starts the snapshot scheduler. In case
there is a volume snapshot policy which exists for a volume which does
not exist, it can cause SQL constraint issue and cause the management
server to break from starting its various components and cause HTTP 503
error.

Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>

* remove schedule on missing volume

---------

Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>

Author: yadvr

server/src/main/java/com/cloud/storage/snapshot/SnapshotSchedulerImpl.java

@@ -415,6 +415,13 @@ public Date scheduleNextSnapshotJob(final SnapshotPolicyVO policy) {
         if (policyId == Snapshot.MANUAL_POLICY_ID) {
             return null;
         }
+
+        if (_volsDao.findById(policy.getVolumeId()) == null) {
+            s_logger.warn("Found snapshot policy ID: " + policyId + " for volume ID: " + policy.getVolumeId() + " that does not exist or has been removed");
+            removeSchedule(policy.getVolumeId(), policy.getId());
+            return null;
+        }
+
         final Date nextSnapshotTimestamp = getNextScheduledTime(policyId, _currentTimestamp);
         SnapshotScheduleVO spstSchedVO = _snapshotScheduleDao.findOneByVolumePolicy(policy.getVolumeId(), policy.getId());
         if (spstSchedVO == null) {