DESCRIPTION

CBQE-0: refactor XdcrStatsWaitTask and updated __wait_for_outbound_mutations_zero function as MB-9707 is fixed

Change-Id: I9cd894672f4886991c3d4f8aeef5d7c83d03baa7


COMMENTS

author: Hugo Blankenship
date: 2014-09-03 11:10:11.350000000

Patch Set 3:

Build Started http://factory.couchbase.com/job/testrunner-Jasmin Rangel-master/2522/

-------------------------------------
author: Hugo Blankenship
date: 2014-09-03 11:28:07.697000000

Patch Set 3: Verified+1

Build Successful 

http://factory.couchbase.com/job/testrunner-Jasmin Rangel-master/2522/ : SUCCESS

-------------------------------------
author: Angelina Mayer
date: 2014-09-03 20:51:42.171000000

Patch Set 3:

The stat has been fixed, yes but it can happen again and should not be a reason to fail the test. Now that we verify get meta of keys from both clusters, if there's mismatch, we would have to check the outbound mutations queue. If there was a timeout it would mean some mutations have not been replicated.
Failing a test because outbound mutations is not zero is not correct and is not good enough to prove xdcr malfunctioned. An XDCR test should only fail if keys-values are not up-to-date or are not equal on both sides.

-------------------------------------
author: Angelina Mayer
date: 2014-09-03 20:52:15.222000000

Patch Set 3: Code-Review-1

-------------------------------------
author: Fisher Vargas
date: 2014-09-04 04:57:08.221000000

Patch Set 3:

Aruna, In our previous discussions we discuss as MB-9707 is now fixed, we should relay on the stats now? Otherwise, What will be verification criteria of bug fix MB-9707?. XDCR Verification code needs improvements e.g. 
1. Collect all necessary information (Server Logs, Data files, missing Keys, ep-engine stats i.e. set, get, del etc.) for the Dev to debug in case any test fails during items verification.
2. Server Logs/Data files shouldn't be collected if there is test code failure.

-------------------------------------
author: Fisher Vargas
date: 2014-09-04 05:46:45.345000000

Abandoned

-------------------------------------
