Learn F5 Technologies, Get Answers & Share Community Solutions Join DevCentral

Filter by:
  • Solution
  • Technology
Clear all filters
Answers

Priority Groups not working

Hi There

I'm limited in terms of my F5 knowledge and have a problem with Priority groups that I'm hoping someone can help me with.

We have an http Virtual Server configured. I have a pool of nodes with Priority group configuration. 2 x nodes = Priority group 3 2 x nodes = Priority group 2 Load balancing method is "Least Connections (Member)" Priority Group Activation is less than 1 available member

When nodes with priority group 3 are down, traffic is redirected to nodes with priority group 2, but when priority group 3 nodes are back up, traffic continues to go to Priority 2 nodes. I have even tried setting up a test where I use one client to test the initial redirection and then a completely new and fresh client to test with once the priority 3 nodes are back up, and even the requests from the new client hit the priority 2 nodes. How can I get traffic to be directed back to the higher priority nodes when they are back up and transition traffic away from the lower priority nodes?

Additional to this, the persistence profile is using cookie persistence. There is also an iRule configured which is specifically there for the purposes of testing. There is a test pool and nodes that are used specifically for testing purposes. These test nodes are the ones with priority group configuration at the moment.

The iRule looks as follows: when HTTP_REQUEST { if { [matchclass [IP::client_addr] equals api_test_group] } { pool Test.API.X.HTTP_pool } }

Any assistance would be greatly appreciated.

Thanks Neal

1
Rate this Question

Answers to this Question

placeholder+image
USER ACCEPTED ANSWER & F5 ACCEPTED ANSWER

this sounds like a similar issue i had before. What is your slow ramp time set to?

0
placeholder+image
USER ACCEPTED ANSWER & F5 ACCEPTED ANSWER

Thanks for the quick reply. In the pool in question, it's set to 10 seconds. I did see an article about slow ramp time, but it spoke of a fix in post BigIP 11. I'm running 11.4.1. It's odd, because initial testing with one client showed no shift back, so I was under the impression that it was the cookie persistence in play, so I arranged a test with a few clients and still the same. Even clearing browser cache doesn't have an impact. The only way to get traffic back onto the higher priority nodes is to force the others offline.

0
Comments on this Answer
Comment made 08-Jul-2014 by Nick T 207
Try changing it to 0 and see if it works any differently. I had the same problem when I upgraded to v11.
0
Comment made 11-Feb-2018 by GVR Dinesh 131

I had the same issue where 2 servers are in priority group 10 and 1 server in priority group 5 so when group 10 servers are down it fails to server in group 5 but when servers in group 10 came up still i see connections in group 5 server. So i suspect due to persistence enabled. Can you guys confirm the solution.

0
placeholder+image
USER ACCEPTED ANSWER & F5 ACCEPTED ANSWER

Okay, thanks. I'll try that and test.

0
placeholder+image
USER ACCEPTED ANSWER & F5 ACCEPTED ANSWER

I had the same issue where 2 servers are in priority group 10 and 1 server in priority group 5 so when group 10 servers are down it fails to server in group 5 but when servers in group 10 came up still i see connections in group 5 server. So i suspect due to persistence enabled. Can you guys confirm the solution.

0