Hi there. I’m trying to create a Wordpress cluster using this tutorial:
http://rancher.com/building-microservices-with-docker-on-the-new-rancher-beta/
I’m all the way at the end and don’t see anywhere in the UI to add the health check functionality depicted in the article. Just wondering if that functionality has gone away.
Thanks much for any insight.
Regards,
Kory
The trouble with blog posts is they don’t automatically keep themselves up to date 
In the current releases you define the health check policy on the target service (WordPress). This works for service discovery in general, failing services are not advertised in DNS and will get replaced if they don’t become healthy again. And when you balance to a service the same health check is used to determine which containers are healthy to send traffic to. So it is not necessary to define this twice anymore.
Thanks, Vincent, I appreciate your reply. I was wondering if the lack of the health check was preventing the containers from building as I never got to that point. But, yes, indeed, the posts don’t keep themselves up to date as they really should. 
I’ll go back to the drawing board and see if I can get this created.
Thanks again!
Well, I got down to the WP section. The service keeps bouncing across the four containers I’ve set up and am seeing this in the log:
Invalid Date Invalid DateMount failed. Please check the log file for more details.
Invalid Date Invalid Date=> Checking if I can reach GlusterFS node 10.42.52.62 …
Invalid Date Invalid Date=> GlusterFS node 10.42.52.62 is alive
Invalid Date Invalid Date=> Mounting GlusterFS volume ranchervol from GlusterFS node 10.42.52.62 …
Invalid Date Invalid DateMount failed. Please check the log file for more details.
Invalid Date Invalid Date=> Checking if I can reach GlusterFS node 10.42.126.92 …
Joins Gluster then disconnects then joins and disconnects.
Anyone have any ideas?
Regards,
Kory
The same tutorial wont work for me… Using the xtradb commented in the comments I was able to get all up including gluster, but WP just wont show anything…blacnk screen… any url blank… no 404 or anything either…
Looking for an uptodate tutorial if anyone has one…
Check the logs for the WordPress container(s), and the Network and Console tabs of the browser developer tools while loading the page.
Nothing anywhere… Nothing in the logs and nothing in the browser/network/inspection…
I’m noticing some weird behavior on LB, not sure if it can be related…
Created a single mysql container exposing 3306:3306.
— I can access the hostip:3306
Create a LB instance listening on 3307 connected to container mysql port 3306
— No response when trying LBHost:3306 (its same host for examples sake)
Create a LB instance listening on 3306 connected to container mysql
— No response when trying LBHost:3306 (its same host for examples sake)
I notice that when I configure the LB to listen on 1 port on host, and pass it to another in container it doesnt “stick”… lb config shows the “public” port twice… ex: I want public port to be 3307, LB option after saving shows Host port: 3307, Container port: 3307…
So I removed the host port on the container, and left it as 3306:3306 in the LB, still no go…
I can only connect directly to the container if I expose it on the host…whenever I try to insert a LB it fails.
Any relation?
If you expose the port of the mysql container on the host, you won’t be able to access it from a LB on the same host.
Where are you seeing these “ports”. Are you looking at the haproxy.cfg or just relying on the UI or the ports on docker?
Can you share your docker-compose.yml file? If you created on the UI, you can go to the stack and look click on this button to find it.
This is my latest attempts…
mysql2:
ports:
- 3333:3306/tcp
restart: always
environment:
MYSQL_ROOT_PASSWORD: test
tty: true
image: mysql
stdin_open: true
mysql:
ports:
- 3306:3306/tcp
restart: always
environment:
MYSQL_ROOT_PASSWORD: test
tty: true
image: mysql
privileged: true
volumes:
- /var/lib/mysql:/var/lib/mysql
stdin_open: true
In this case, I can reach host:3333 through mysql client, but not host:3306… weird…
Now adding a LB:
mysql2-lb:
ports:
- 3336:3306
restart: always
tty: true
image: rancher/load-balancer-service
links:
- mysql2:mysql2
stdin_open: true
mysql2:
ports:
- 3333:3306/tcp
restart: always
environment:
MYSQL_ROOT_PASSWORD: mestre
tty: true
image: mysql
stdin_open: true
mysql:
ports:
- 3306:3306/tcp
restart: always
environment:
MYSQL_ROOT_PASSWORD: mestre
tty: true
image: mysql
privileged: true
volumes:
- /var/lib/mysql:/var/lib/mysql
stdin_open: true
Accessing mysql directly on host:3333 works, but accessing host:3336 (mysqlLB->mysql) doesnt connect…
Also, visually when I click on LB_Agent its reporting the wrong port (that had thrown me off but apparently its only visual)
Ports
State IP Address Public (on Host) Private (in Container) Protocol
ACTIVE 3336 3336 TCP
I hope this can help somewhat?
(In all cases my initial container with exposed port 3306 isnt working… I think it stopped working when I added an LB, but its weird nonetheless? unless there is some sort of port reservation on the hosts 3306?)
Just an FYI - im using only one host, so when I say host:xxxx all containers are on the same host…
There appears to be a UI bug, which I also hit and filed, where we are showing the incorrect private port. But you can check the haproxy.cfg for the load balancer to make sure it was configured correctly. http://docs.rancher.com/rancher/faqs/troubleshooting/#load-balancer
https://github.com/rancher/rancher/issues/2272
As for the docker-compose.yml, for mysql2-lb, I’m not too sure why the LB wouldn’t work. Could you try using a different container like nginx instead of mysql?