Forum Discussion
CirrusAll client connection are getting closed when one of the pool member is down with MBLB profile on SIP.
We have a virtual server with 2 pool members and a MBLB profile assocaited with it
VIRTUAL SERVER:
============
ltm virtual /Common/SIP-TLS {
destination /Common/192.168.3.57:5061
ip-protocol tcp
mask 255.255.255.255
pool /Common/SIP_TCP_POOL_MBLB
/Common/mymblb { }
/Common/sip { }
/Common/tcp { }
}
rules {
/Common/TCP_Pool_rule
}
snat automap
translate-address enabled
translate-port enabled
vlans-disabled
}
MBLB profile:
=========
ltm profile mblb /Common/mymblb {
defaults-from /Common/mblb
ingress-high 10000
ingress-low 9000
isolate-abort enabled
isolate-client enabled
isolate-expire enabled
isolate-server disbaled
min-conn 4
tag-ttl 60
}
POOL:
=====
ltm pool /Common/SIP_TCP_POOL_MBLB {
members {
/Common/192.168.101.63:5090 {
address 192.168.101.63
}
/Common/192.168.101.64:5090 {
address 192.168.101.64
}
}
monitor /Common/gateway_icmp
service-down-action reset
}
Whenever a pool member (server) health is going bad or pool member (server) is down F5 is bringing down all client connections connected to that virtual server.
Are there any configuration or iRule to bring down only a set of client connection that are connected to a Server (pool member which is down) instead of F5 taking down all the client connections?
3 Replies
hoolioCirrostratus
I would expect that only the client connections who have been load balanced to the down pool member would be affected. If you're seeing all client connections reset when one pool member is marked down by a monitor, I would open a case with F5 Support.
SOL135 - Information required when opening a support case for BIG-IP LTM or GTM
https://support.f5.com/kb/en-us/solutions/public/0000/100/sol135.html
Aaron
jsikorski_47403Nimbostratus
There's a typo in the MBLB profile you posted:
isolate-server disbaled
Did that come from your actual config?- linjing
Employee
old post, but at last is this caused by : service-down-action reset ?
