problem with my sorry server

REDES_STI_UCHILE

Hello:
I have two service-group,(servers X and sorry servers). When my primary service-group fail (all servers down) sorry server respond in other service-group (this step is okay) but when primary service-group came back, I (have) need to wait about 8 minutes (that time is a configuracion for sticky in primary service-group). There are some configuration for not consider "the sticky (template source IP)" in the service-group backup (sorry) and could connect immediately when my service-group primary came back.

this is a example of my configuration:
-----------------------------------------------------------
Primary service-group:
*************************
Aflex

when LB_FAILED {
pool cluster-sorry
}
**********************

slb template persist source-ip sticky_prueba-saamatricula
match-type service-group
timeout 3

slb virtual-server VIP_cluster-prueba-saamatricula 172.X.X.X
extended-stats
port 443 tcp
name XXXXX
access-list 100 source-nat-pool source_nat
service-group cluster-prueba-saamatricula
extended-stats
ha-conn-mirror
template persist source-ip sticky_prueba-saamatricula
aflex SORRY

-----------------------------------------------------------------------
sorry server

slb virtual-server VIP_cluster-sorry 172.X.X.X
extended-stats
port 443 tcp
access-list 100 source-nat-pool source_nat
service-group cluster-sorry
extended-stats
ha-conn-mirror

I need I help with that please.
regards.

dshin

Looking at your configuration, the traffic will remain to take the "Primary Service-Group" path as the aFlex is bound on that VIP. The aFlex will failover to the “Sorry Server” and the template persist source-ip feature will also take in effect. The Source-IP config might be causing the 8 mins persistence problem(default is 5 mins). You can try to adjust the timeout to 1 minute and see if the failover delay makes a difference.

slb template persist source-ip sip
timeout 1

Genard