Incorrect NonIO is reported by cbstats all

Problem

The number of nonIO threads reported by cbstats all is incorrect when it's changed via another bucket. This lead the user to believe that it was a bucket level setting and not a global one.

Steps to reproduce

1. Setup a 1 node cluster with just the Data service and two buckets
2. Change the number of nonIO threads via bucket B

   /opt/couchbase/bin/cbepctl localhost:11210 -b B -u Administrator -p password set flush_param num_nonio_threads 8

   setting param: num_nonio_threads 8

   set num_nonio_threads to 8

3. Check the if it has changed for bucket A and B using cbstats all

   [vagrant@node1-cb710-centos7 ~]$ /opt/couchbase/bin/cbstats localhost:11210 -b A -u Administrator -p password all | grep nonio

   ep_num_nonio_threads: 2

   [vagrant@node1-cb710-centos7 ~]$ /opt/couchbase/bin/cbstats localhost:11210 -b B -u Administrator -p password all | grep nonio

   ep_num_nonio_threads: 8

Note

cbstats workload does report the correct value:

 /opt/couchbase/bin/cbstats localhost:11210 -b A -u Administrator -p password workload 

 ep_workload:LowPrioQ_AuxIO:InQsize:   2

 ep_workload:LowPrioQ_AuxIO:OutQsize:  0

 ep_workload:LowPrioQ_NonIO:InQsize:   2072

 ep_workload:LowPrioQ_NonIO:OutQsize:  0

 ep_workload:LowPrioQ_Reader:InQsize:  2

 ep_workload:LowPrioQ_Reader:OutQsize: 0

 ep_workload:LowPrioQ_Writer:InQsize:  4

 ep_workload:LowPrioQ_Writer:OutQsize: 0

 ep_workload:num_auxio:                2

 ep_workload:num_nonio:                8

 ep_workload:num_readers:              4

 ep_workload:num_shards:               1

 ep_workload:num_sleepers:             18

 ep_workload:num_writers:              4

 ep_workload:ready_tasks:              0