rebalancing a cluster from 3 nodes to 2 nodes, keeps many vbuckets on dead state.

Created an attachment (id=740)
logs

The first node on the cluster was 10.1.4.27. Adding 10.1.4.25, 10.1.4.26, than removing 10.1.4.26.

[root@Config131VM0 vizBucket]# ./vbucketctl 10.1.4.25:11210 list | awk '

{print $3}' | sort | uniq -c
512 active
206 dead
235 replica
[root@Config131VM0 vizBucket]# ./vbucketctl 10.1.4.27:11210 list | awk '{print $3}

' | sort | uniq -c
512 active
16 dead
251 replica
[root@Config131VM0 vizBucket]# ./vbucketctl 10.1.4.25:11210 list | awk '

{print $3}' | sort | uniq -c
512 active
206 dead
235 replica
[root@Config131VM0 vizBucket]# ./vbucketctl 10.1.4.27:11210 list | awk '{print $3}

' | sort | uniq -c
512 active
16 dead
251 replica
[root@Config131VM0 vizBucket]#