DESCRIPTION

CBQE-1518: add test case to verify fix for MB-8825

Change-Id: Ie379090d8c24b41d42d7d9af80d48afe5852a17a


COMMENTS

author: Trevor Santana
date: 2013-11-08 07:33:58.139000000

Patch Set 1: Code-Review+2

-------------------------------------
author: Ronald Acevedo
date: 2013-11-08 16:42:55.457000000

Patch Set 1: Code-Review+1

-------------------------------------
author: Macie Morton
date: 2013-11-14 16:05:30.307000000

Patch Set 1:

(2 comments)

Hi Sangarsh, did you get run the test and it passes?

I am not sure I understand correctly. At verification stage, the stats  you read 
from source cluster ep_num_ops_del_meta_fail and _ep_num_ops_set_meta_res_fail  should not be 0. See my inline comments.

Line:58, pytests/xdcr/xdcr_dataloss.py -> Is it correct here? 

The expected behaviour is all backward mutations from remote back to source should be blocked. In this case, all returning mutations from remote should fail conflict resolution, both setWithMeta and delWithMeta.

If I understand correctly, stat_ep_num_ops_del_meta_fail means # of failures, thus it is expected to be # of all deletions sent from remote to source, not 0.

Line:60, pytests/xdcr/xdcr_dataloss.py -> Same as above.

stat_ep_num_ops_set_meta_res_fail should be # of all setWithMeta from remote.

-------------------------------------
author: Macie Morton
date: 2013-11-14 16:05:47.673000000

Patch Set 1: Code-Review-1

-------------------------------------
author: Macie Morton
date: 2013-11-14 16:08:36.654000000

Patch Set 1:

Also, please see my comment at MB-8825 how we use these two stats to debug.

-------------------------------------
author: Fisher Vargas
date: 2013-11-15 16:57:57.351000000

Patch Set 2: Patch Set 1 was rebased

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