Forum Discussion
LB_FAILED triggered when there are available members in the pool
Before I resorted to an irule, I tried using the priority group method which in theory worked... however we had many complaints of customers getting sent to the lower priority group w/ a splash page when the higher priority group was not full. I had evidence of this in stats & opened a case with F5 which never actually got resolved & my rep suggested I use an irule instead.
Here is my current iRule:
when RULE_INIT {
Set an HTML response to sent to clients who make a request while the VIP is over the max connection count
set ::html_content "Insert busy splash page html content here"
}
when LB_FAILED {
HTTP::respond 200 content $::html_content
}
As a work around I've doubled the allowed connection limits in the pools so the iRule will allow it to get to the actual limits I want, however I don't want it to all of a sudden start allowing the entire 2000 connections in on 1 node. Other than that, the rule works amazing & haven't had any other customer complaints. Any help would be appreciated!
- Richard__HarlanHistoric F5 AccountIf you are running a LTM with CMP you could be running into the issue where connection limits are spread across all the TMM. So for each TMM the connection limit per pool/member is 1/n the total count. There could be cases were if too many connection cluster to one TMM then the TMM in question will run into the connection limit issue. You could also run into this issue if the VIP in question has been demoted from CMP.
- nitassEmployeespark has some suggestion in the following discussion topic. anyway, i am not sure if it is practical in case of http.
Recent Discussions
Related Content
* Getting Started on DevCentral
* Community Guidelines
* Community Terms of Use / EULA
* Community Ranking Explained
* Community Resources
* Contact the DevCentral Team
* Update MFA on account.f5.com