Description of problem:
Logging stack started but logging-es logs have:

   * Try running sgadmin.sh with -icl and -nhnv (If thats works you need to check your clustername as well as hostnames in your SSL certificates)
[2017-06-08 14:04:59,526][ERROR][com.floragunn.searchguard.auth.BackendRegistry] Not yet initialized (you may need to run sgadmin)

In indicies check we see:


green  open   .searchguard.logging-es-3ec2rtmj-2-d6wzd                                               1   2          4            0    138.4kb         46.1kb
red    open   .searchguard.logging-es-3ec2rtmj-1-evitf                                               1   2                                            

green  open   .searchguard.logging-es-l5nl20ab-1-emkhw                                               1   2          4            0     97.2kb         32.4kb
red    open   .searchguard.logging-es-l5nl20ab-1-88m47                                               1   2                                            


Version-Release number of selected component (if applicable):
v3.5

Jeff Cantril indicates PR might backport a fix to this issue:
https://github.com/openshift/origin-aggregated-logging/pull/460


Potential workaround: 
scaling down elasticsearch cluster to 1 node; delete all .searchguard indicies; redeploy.
Is this a viable workaround? Can we clarify if "redeploy" means run the deployer template with MODE=redeploy or if it means `oc rollout latest` or something similar. Or do we just scale up to 3?