OS Patching process and Lost stats

I’m interested in how other people are doing the patching process. We are running version 5.5.4 EE with 4 data nodes and 2 index\query nodes on a Red Hat 7.

  1. Currently we fail one node out rebalance the cluster.
  2. OS patch the node that has been removed + reboot.
  3. Add the node back into the cluster and rebalance again.
    … and repeat till the cluster is upgraded.

…but I have noticed that all my stats get wiped ???

Now in the below link
it says the stats are held in /opt/couchbase/var/lib/couchbase/stats

… so should I backup this directory before we remove the node from the cluster then restore in when we add it back in again?

I’m guessing when we add the node back into the cluster its wiping all the stats from the directory.

So I’m interested to know if there is a better\recommend way to patch the nodes?
thanks in advanced.