DESCRIPTION

2i - Add a check to verify cluster status in test_rebalance_in

Before performing after recovery tasks, there is no check if the
cluster is in healthy state. Check if all the nodes in the cluster
are healthy before doing any post recovery task.

Change Query Definition to remove dead code and recategorize
queries into various groups.

Change-Id: I7eed5028823760de35c12514acdba97dedfbc8ae
Partial-Bug: CBQE-3739


COMMENTS

author: Coby Mathis
date: 2016-12-19 10:41:27.749000000

Uploaded patch set 1.

-------------------------------------
author: Lucille Zavala
date: 2016-12-19 10:58:47.435000000

Patch Set 1:

(3 comments)

Line:79, pytests/2i/recovery_2i.py -> Don't we need to verify that cluster is healthy in rebalance out scenario's too?

Line:232, pytests/2i/recovery_2i.py -> Intentional?

Line:393, pytests/2i/recovery_2i.py -> Intentional? If so, why?

-------------------------------------
author: Coby Mathis
date: 2016-12-19 11:07:15.221000000

Patch Set 1:

(3 comments)

Line:79, pytests/2i/recovery_2i.py -> This should be there in all recovery scenarios. I am starting with rebalance_in.

Line:232, pytests/2i/recovery_2i.py -> Yes. With current design, replica indexes are being created in setup itself because of which, this method fails creating indexes since duplicate index already exists.

Line:393, pytests/2i/recovery_2i.py -> Yes. Same reason as above

-------------------------------------
author: Lucille Zavala
date: 2016-12-19 11:08:54.600000000

Patch Set 1: Code-Review+1

(1 comment)

Line:232, pytests/2i/recovery_2i.py -> Then just delete the line instead of commenting it out? Keeps the code cleaner.

-------------------------------------
author: Adelaide Rodgers
date: 2016-12-20 05:10:49.985000000

Patch Set 1: Code-Review-1

(2 comments)

Line:765, pytests/2i/base_2i.py -> this will exit the for loop and set is_cluster_healthy to True which will exit the while loop - please correct the logic

Line:3402, lib/membase/api/rest_client.py -> nitpick - would sound better as "fix for MB-20758" or "workaround for MB-20758"

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