Forum Discussion

vladimir_klepc1's avatar
vladimir_klepc1
Icon for Nimbostratus rankNimbostratus
Apr 23, 2016

Strange round robin v12

Hi. I use very simple load balansing method - round robin. Pool consists of three members. I expected to get sequence 1,2,3,1,2,3,1,2,3 or something similar but got 3,1,3,2,2,1,2,1,3. I use virtual edition version 12.

 

root@(bigip1)(cfg-sync In Sync)(Active)(/Common)(tmos) list ltm virtual ltm virtual test_ssh { destination 172.27.34.11:ssh ip-protocol tcp mask 255.255.255.255 pool ssh_pool profiles { tcp { } } source 0.0.0.0/0 source-address-translation { type automap } vs-index 3

 

ltm pool ssh_pool { members { 172.27.142.101:ssh { address 172.27.142.101 session monitor-enabled state up } 172.27.142.102:ssh { address 172.27.142.102 session monitor-enabled state up } 172.27.142.103:ssh { address 172.27.142.103 session monitor-enabled state up } } monitor gateway_icmp }

 

application delivery
LTM
  • vladimir_klepc1's avatar
    vladimir_klepc1
    Icon for Nimbostratus rankNimbostratus
    Apr 24, 2016
    I checked behavior on version 11.5. Result the same. Not either CMP (use VE) or irules in use. All sessions during test in active state. Any ideas why round robin doesn't balance in sequence? /Common/ssh_pool 172.27.142.101 22 /Common/ssh_pool 172.27.142.102 22 /Common/ssh_pool 172.27.142.101 22 /Common/ssh_pool 172.27.142.101 22 /Common/ssh_pool 172.27.142.102 22 /Common/ssh_pool 172.27.142.102 22 /Common/ssh_pool 172.27.142.103 22 /Common/ssh_pool 172.27.142.101 22
  • vladimir_klepc1's avatar
    vladimir_klepc1
    Icon for Nimbostratus rankNimbostratus
    Apr 24, 2016

    Found the answer. VE also uses a TMM whis is a multi-threaded process. When I disable multi-threading on VS with command "modify ltm virtual test_ssh cmp-enabled no" or deacrease vCPU to one, load distribution become as a expected: node №1, node №2, node №3, node №1, node №2, node №3 and so on.