ocf_heartbeat_redis resource not able to point the slave to the master on failback

the redis resource is not able to point the slave node to the new master redis server on a failback operation. The other resources are back up correctly

Slave node logs: (note the line `Setting master to 'no-such-master'`)
`Sep 29 19:07:44 IPaddr2(virtual_public_ip)[13245]: INFO: Adding inet address 192.168.56.122/24 with broadcast address 192.168.56.255 to device enp0s9
Sep 29 19:07:44 IPaddr2(virtual_public_ip)[13245]: INFO: Bringing device enp0s9 up
Sep 29 19:07:44 IPaddr2(virtual_public_ip)[13245]: INFO: /usr/lib/heartbeat/send_arp -i 200 -r 5 -p /var/run/resource-agents/send_arp-192.168.56.122 enp0s9 192.168.56.122 auto not_used not_used
Sep 29 19:07:45 IPaddr2(virtual_public_ip)[13245]: INFO:
Sep 29 19:08:11 IPaddr2(virtual_public_ip)[13531]: INFO: IP status = ok, IP_CIP=
Sep 29 19:08:12 redis(p_redis)[13631]: INFO: demote: Setting master to 'no-such-master'
Sep 29 19:08:14 redis(p_redis)[13736]: INFO: monitor: Slave mode link has not yet been established (link=down)
Sep 29 19:08:14 redis(p_redis)[13736]: INFO: demote: Setting master to 'no-such-master'`

Redis resource:
`primitive p_redis redis \
op monitor timeout=60 interval=45 depth=0 \
op monitor role=Master timeout=60 interval=20 depth=0 \
op monitor role=Slave timeout=60 interval=60 depth=0`

clone resource:
`ms redis_clone p_redis \
meta notify=true interleave=true`

I also have a constant to start the virtual ip first of the redis resource:
`colocation ip-on-redis inf: virtual_public_ip redis_clone:Master`

Start with the ocf_heartbeat_redis resource agent and review how the demote and monitor operations handle the master address during failback. Reproduce the reported failback using the p_redis and redis_clone configuration, then verify that the former slave is pointed to the new master instead of no-such-master.

Written by the indexing model from the issue text.