XI52 188.8.131.52 Firmware, Weblogic back-end
This should be easy, but for some strange reason, we cannot seem to use a Load Balancer Group as a back end for an MPG.
Say our LBG has 2 members and is named MY_LBG. The members' are defined by IP, and the mapped server port is set to 0. We set the static back end in the MPG to "http://MY_LBG:8011", and BOTH servers are listening on port 8011. In this setup, it says it fails to make the back-side connection. So, I thought I had the port setup backwards, and changed the mapped server port on each member to 8011, and changed the static back end in the MPG to "http://MY_LBG". Again, it says it fails to make the back-side connection.
If I set the static back end to any of the members directly, like "http://IP_OF_MEMBER_1:8011" or "http://IP_OF_MEMBER_2:8011", it works just fine.
So, I'm thinking I'm missing a radio button somewhere.
Here's what we have set:
Retrieve Workload Management Information: off
Damp Time: 120
Do Not Pass Down State: off
Try Every Server Before Failing: on
Masquerade as group name: on
HEALTH *** Not enabled
Insertion Cookie Name: DPJSESSIONID
Insertion Path: /
Insertion Domain: .ourcompany.com
Override Websphere Cell Session Affinity: on
Monitored Cookies: JSESSIONID
MEMBERS (2 identically setup except IP address)
Mapped Server Port: 0 (or the actual of 8011, neither worked)
Health Port: 0 ( Health Checking is off right now )
Pinned topic Need Help with Load Balancer Group as MPG Backend
Answered question This question has been answered.
Unanswered question This question has not been answered yet.
Updated on 2013-02-25T20:52:09Z at 2013-02-25T20:52:09Z by SystemAdmin
SystemAdmin 110000D4XK6772 Posts
Re: Need Help with Load Balancer Group as MPG Backend2013-02-25T20:52:09ZThis is the accepted answer. This is the accepted answer.
- msiebler 2700005RPQ
We did solve it, but didn't "solve" anything, in the sense of understanding it.
We dropped the LBG and re-built it. It is working now. In my haste to rebuild it, I gave the new one a really short name. The first one had a pretty long name, like 25 to 30 characters.
The one that is now working is 8 characters long.
Is there anything in anyone's experience where the length of the name of a load balancer group affected its working state?