Update README.md

Author: doublez13

README.md

@@ -29,7 +29,7 @@ As Traefik needs to be aware of containers stopping and starting, it uses the do
 One way to mitigate the risks is to wrap the docker socket with a proxy that only allows a subset of commands, and then route all docker socket communication through that. For this, we'll be using the popular [docker-socket-proxy](https://github.com/Tecnativa/docker-socket-proxy). Using docker-socket-proxy, by default most requests are blocked. This includes all non-GET requests. For Traefik to function, we'll need to open up a few additional docker socket endpoints. Finally, we'll make the network that the docker socket proxy lives on internal, and attach only the Traefik container to it.  
 
 ## Keepalived:
-Docker swarm may be able to move your containers to another node if a problem is detected, but this is useless if any DNS entries point to the problem node. When exposing a port in Docker swarm, that port is exposed on all nodes. An external load balancer is often used to balace the requests among the swarm nodes. One benefit is that a load balancer can often check the health of a node, and stop sending it requests if it detects a problem. But again, the load balancer becomes a single point of failure. In this example, we'll be using VRRP ([Virtual Router Redundancy Protocol](https://en.wikipedia.org/wiki/Virtual_Router_Redundancy_Protocol)). VRRP allows all nodes in the cluster to share one (or more) IP addresses. This "virtual address" can live on it's own interface, or share the interface associated with the node's primary address. Only one node in the cluster has the IP address at any one time. If that node goes offline, the IP address seamlessly migrates to another node in a matter of seconds. This example uses keepalived to implement VRRP. In the [example coniguration provided](https://github.com/doublez13/docker-swarm-example-setup/blob/master/keepalived/keepalived.conf), keepalived monitors the docker service.
+Docker swarm may be able to move your containers to another node if a problem is detected, but this is useless if any DNS entries point to the problem node. When exposing a port in Docker swarm, that port is exposed on all nodes. An external load balancer is often used to balace the requests among the swarm nodes. One benefit is that a load balancer can often check the health of a node, and stop sending it requests if it detects a problem. But again, the load balancer becomes a single point of failure. In this example, we'll be using VRRP ([Virtual Router Redundancy Protocol](https://en.wikipedia.org/wiki/Virtual_Router_Redundancy_Protocol)). VRRP allows all nodes in the cluster to share one (or more) IP addresses. This "virtual address" can live on it's own interface, or share the interface associated with the node's primary address. Only one node in the cluster has the IP address at any one time. If that node goes offline, the IP address seamlessly migrates to another node in a matter of seconds. This example uses keepalived to implement VRRP. In the [example configuration provided](https://github.com/doublez13/docker-swarm-example-setup/blob/master/keepalived/keepalived.conf), keepalived monitors the docker service, and also gives preference to the host running the Traefik instance.
 
 ## NFS:
 There are many third party Docker volume drivers out there. For this setup, I'm just using the local driver to NFS mount docker volumes. I have seen some mixed feedback on this, but my understanding is as follows. Do **NOT** NFS mount the docker volumes directory (/var/lib/docker/volumes). This is just asking for trouble when two different docker daemons think they have ownership. However, I have seen no problems using the local driver with type set to NFS. Docker automatically mounts and unmounts the volumes inside the docker volumes directory. Check out the wordpress compose file for an example.