Description of problem:

When restarting atomic-openshift-master-api services on all masters 1 or all will fail with the following error: 


Dec 30 13:06:19 master3.bender.com atomic-openshift-master-api[57882]: F1230 13:06:19.611303   57882 controller.go:83] Unable to perform initial IP allocation check: unable to persist the updated service IP allocations: serviceipallocation "" cannot be updated: the provided resource version does not match


Version-Release number of selected component (if applicable):
3,1

How reproducible: 80%


Steps to Reproduce:
1. Reboot all masters at the same time
 or restart all atomic-openshift-master-api.services together

# ansible masters -a "systemctl restart atomic-openshift-master-api"

Actual results:
One fails most of the time with the error below

Expected results:

All to start up with out error


Additional info:


master3.bender.com | FAILED | rc=3 >>
* atomic-openshift-master-api.service - Atomic OpenShift Master API
   Loaded: loaded (/usr/lib/systemd/system/atomic-openshift-master-api.service; enabled; vendor preset: disabled)
   Active: failed (Result: exit-code) since Wed 2015-12-30 13:16:42 EST; 2min 3s ago
     Docs: https://github.com/openshift/origin
  Process: 59189 ExecStart=/usr/bin/openshift start master api --config=${CONFIG_FILE} $OPTIONS (code=exited, status=255)
 Main PID: 59189 (code=exited, status=255)


Dec 30 13:16:41 master3.bender.com systemd[1]: Starting Atomic OpenShift Master API...
Dec 30 13:16:41 master3.bender.com atomic-openshift-master-api[59189]: I1230 13:16:41.729040   59189 start_api.go:96] Using a masterIP override "192.168.110.12"
Dec 30 13:16:41 master3.bender.com atomic-openshift-master-api[59189]: I1230 13:16:41.729179   59189 start_api.go:102] Using a listen address override "0.0.0.0:8443"
Dec 30 13:16:42 master3.bender.com atomic-openshift-master-api[59189]: I1230 13:16:42.498804   59189 plugins.go:71] No cloud provider specified.
Dec 30 13:16:42 master3.bender.com atomic-openshift-master-api[59189]: I1230 13:16:42.626631   59189 start_master.go:388] Starting master on 0.0.0.0:8443 (v3.1.0.4-16-g112fcc4)
Dec 30 13:16:42 master3.bender.com atomic-openshift-master-api[59189]: I1230 13:16:42.626673   59189 start_master.go:389] Public master address is https://openshift-cluster.bender.com:8443
Dec 30 13:16:42 master3.bender.com atomic-openshift-master-api[59189]: I1230 13:16:42.626690   59189 start_master.go:393] Using images from "openshift3/ose-<component>:v3.1.0.4"
Dec 30 13:16:42 master3.bender.com atomic-openshift-master-api[59189]: I1230 13:16:42.650845   59189 run_components.go:183] Using default project node label selector:
Dec 30 13:16:42 master3.bender.com atomic-openshift-master-api[59189]: I1230 13:16:42.814940   59189 master.go:369] Setting master service IP to "172.30.0.1" (read-write).
Dec 30 13:16:42 master3.bender.com atomic-openshift-master-api[59189]: F1230 13:16:42.932154   59189 controller.go:83] Unable to perform initial IP allocation check: unable to persist the updated service IP allocations: serviceipallocation "" cannot be updated: the provid
Dec 30 13:16:42 master3.bender.com systemd[1]: atomic-openshift-master-api.service: main process exited, code=exited, status=255/n/a
Dec 30 13:16:42 master3.bender.com systemd[1]: Failed to start Atomic OpenShift Master API.
Dec 30 13:16:42 master3.bender.com systemd[1]: Unit atomic-openshift-master-api.service entered failed state.
Dec 30 13:16:42 master3.bender.com systemd[1]: atomic-openshift-master-api.service failed.