Merge pull request #1529 from mike-tutkowski/marvin_replace_sleep

Marvin: Replace a timer.sleep(30) with pulling logichttps://issues.apache.org/jira/browse/CLOUDSTACK-9374

From the ticket:

In the base.py file, there is a Host class with a delete instance method.

This method first attempts to transition the host into the maintenance resource state.

The first step in this process is to transition the host into the prepare-for-maintenance resource state.

A while later, the host can be transitioned completely into the maintenance resource state.

In an attempt to wait for this transition to occur, the delete method has a timer.sleep(30) call.

The hope is that the host will have transitioned from the prepare-for-maintenance resource state to the maintenance resource state within 30 seconds, but this does not always happen.

We should correct this problem by putting in logic to query the management server for the resource state of the host. If it's in the expected state, move on; else, sleep for a bit and try again (up to a certain limit).

* pr/1529:
  Replace a timer.sleep(30) with pulling logic

Signed-off-by: Will Stevens <williamstevens@gmail.com>

Author: swill

tools/marvin/marvin/codes.py

@@ -115,6 +115,31 @@
 '''
 ALLOCATED = "Allocated"
 
+'''
+Host states
+'''
+HOST_CREATING = "Creating"
+HOST_CONNECTING = "Connecting"
+HOST_UP = "Up"
+HOST_DOWN = "Down"
+HOST_DISCONNECTED = "Disconnected"
+HOST_ALERT = "Alert"
+HOST_REMOVED = "Removed"
+HOST_ERROR = "Error"
+HOST_REBALANCING = "Rebalancing"
+HOST_UNKNOWN = "Unknown"
+
+'''
+Host resource states
+'''
+HOST_RS_CREATING = "Creating"
+HOST_RS_ENABLED = "Enabled"
+HOST_RS_DISABLED = "Disabled"
+HOST_RS_PREPARE_FOR_MAINTENANCE = "PrepareForMaintenance"
+HOST_RS_ERROR_IN_MAINTENANCE = "ErrorInMaintenance"
+HOST_RS_MAINTENANCE = "Maintenance"
+HOST_RS_ERROR = "Error"
+
 '''
 Storage Tags
 '''

tools/marvin/marvin/lib/base.py

@@ -23,9 +23,10 @@
 from marvin.cloudstackAPI import *
 from marvin.codes import (FAILED, FAIL, PASS, RUNNING, STOPPED,
                           STARTING, DESTROYED, EXPUNGING,
-                          STOPPING, BACKED_UP, BACKING_UP)
+                          STOPPING, BACKED_UP, BACKING_UP,
+                          HOST_RS_MAINTENANCE)
 from marvin.cloudstackException import GetDetailExceptionInfo, CloudstackAPIException
-from marvin.lib.utils import validateList, is_server_ssh_ready, random_gen
+from marvin.lib.utils import validateList, is_server_ssh_ready, random_gen, wait_until
 # Import System modules
 import time
 import hashlib
@@ -2459,13 +2460,40 @@ def create(cls, apiclient, cluster, services, zoneid=None, podid=None, hyperviso
                   GetDetailExceptionInfo(e)
             return FAILED
 
+    @staticmethod
+    def _check_resource_state(apiclient, hostid, resourcestate):
+        hosts = Host.list(apiclient, id=hostid, listall=True)
+
+        validationresult = validateList(hosts)
+
+        assert validationresult is not None, "'validationresult' should not be equal to 'None'."
+
+        assert isinstance(validationresult, list), "'validationresult' should be a 'list'."
+
+        assert len(validationresult) == 3, "'validationresult' should be a list with three items in it."
+
+        if validationresult[0] == FAIL:
+            raise Exception("Host list validation failed: %s" % validationresult[2])
+
+        if str(hosts[0].resourcestate).lower().decode("string_escape") == str(resourcestate).lower():
+            return True, None
+
+        return False, "Host is not in the following state: " + str(resourcestate)
+
     def delete(self, apiclient):
         """Delete Host"""
         # Host must be in maintenance mode before deletion
         cmd = prepareHostForMaintenance.prepareHostForMaintenanceCmd()
         cmd.id = self.id
         apiclient.prepareHostForMaintenance(cmd)
-        time.sleep(30)
+
+        retry_interval = 10
+        num_tries = 10
+
+        wait_result, return_val = wait_until(retry_interval, num_tries, Host._check_resource_state, apiclient, self.id, HOST_RS_MAINTENANCE)
+
+        if not wait_result:
+            raise Exception(return_val)
 
         cmd = deleteHost.deleteHostCmd()
         cmd.id = self.id
@@ -2711,7 +2739,7 @@ def getState(cls, apiclient, poolid, state, timeout=600):
                           id=poolid, listAll=True)
                 validationresult = validateList(pools)
                 if validationresult[0] == FAIL:
-                    raise Exception("Host list validation failed: %s" % validationresult[2])
+                    raise Exception("Pool list validation failed: %s" % validationresult[2])
                 elif str(pools[0].state).lower().decode("string_escape") == str(state).lower():
                     returnValue = [PASS, None]
                     break