Forum Discussion

iremhinmhen_bri's avatar
iremhinmhen_bri
Icon for Nimbostratus rankNimbostratus
May 12, 2016

Pool not sending Traffic to a Server

i configured a VS with a default pool, the Pool has 2 servers. using Round Robin LB method, traffic is only going to one server. i did not configure any form of Persistence. and if i disable the server with traffic, the traffic move to the second.

 

how can i resolve this?

 

i have the iRule associated to the VS:

 

brm-iRule

when CLIENT_ACCEPTED {

 

The current pool is derived from the virtual server's "Default Pool"

set current_pool [LB::server pool]

 

Count of reselect attempts

set reselects 0 } when LB_FAILED {

 

Reselect another pool member if the connection fails. Limit reselects to the number of active members. To be active, a member must be enabled with status available or unknown.

if { $reselects < [active_members $current_pool] } { incr reselects log local0.info "Client [IP::client_addr]:[TCP::client_port]: Selected pool member [LB::server addr]:[LB::server port] not responding. Reselect attempt $reselects." LB::mode rr LB::reselect pool $current_pool } else { log local0.error "Client [IP::client_addr]:[TCP::client_port]: Failed to connect after $reselects reselects." } }

 

any contribution to resolving this will be appreciated.

 

application delivery
iRules
LTM

5 Replies

Sort By
  • Kevin_Stewart's avatar
    Kevin_Stewart
    Icon for Employee rankEmployee
    May 12, 2016

    What do your LTM logs say?

     

    Do you have priority activation enabled on the pool?

     

  • Hannes_Rapp's avatar
    Hannes_Rapp
    Icon for Nimbostratus rankNimbostratus
    May 13, 2016
    What is your ratio configuration setting in pool member config and in node config? If you want to get close to 50/50 balancing, they should be equal for both pool members and both nodes.
  • Kevin_Stewart's avatar
    Kevin_Stewart
    Icon for Employee rankEmployee
    May 13, 2016

    As a quick observation I don't see anything in the iRule that might be forcing all traffic to one node, but then maybe some details are missing.

     

    1. In /var/log/ltm do you see any of the log messages form this iRule? And more important, do you see log messages from the LB_FAILED event?

       

    2. You said there's no priority activation on the pool, but are both pool members actively up in accordance with the monitors? Do you have a pool monitor or individual node monitors?

       

    3. Do you have any ratios applied to the pool members?

       

    4. And what happens if you remove the iRule? Do they start load balancing correctly? How are you testing? With a single browser client? Do you close and open the browser between tests?

       

    5. What type of persistence are you using?

       

  • crodriguez's avatar
    crodriguez
    Ret. Employee
    May 13, 2016
    What type of virtual server is this? In other words, what sort of protocol profiles does it have on it? TCP? HTTP? SIP? Anything else?