Skip to content


Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP

CCBC-613: Count data nodes and base ketama/hashing algorithms on that

While vBucket hashing gives us the exact index of the server, ketama
hashing is based on the number of servers actually present. Ensure that
the "number of servers" means the "number of data nodes" rather than
"the number of all servers", since the former will result in some keys
being mapped to non-data nodes

This also changes the remap algorithm in the vbucket code to use only
those nodes which are data nodes as a point of reference, rather than
"all the nodes" (note, this might be superfluous since we never map keys
to nodes which don't have any vbuckets attached to themselves anyway)

Change-Id: Ic507666c09681541f901d430d8ecf9afa6267b15
Tested-by: buildbot <>
Reviewed-by: Subhashni Balakrishnan <>
latest commit 56d99d9f2f
@mnunberg mnunberg authored mnunberg committed
Something went wrong with that request. Please try again.