LB_FAILED triggered for one specific client to server connection
Hello,
We're using BIG-IP LTM, and we're having a problem with one particular IP address causing an LB_FAILED when trying to connect to a pool
Here is relevant part of the redirect rule:
when HTTP_REQUEST {
...
elseif { [HTTP::uri] starts_with "/core/" } {
log local0. "XXX core URL found, sending to my_pool [IP::client_addr]"
pool my_pool
}
...
And here is the maintenance rule:
when LB_FAILED {
LB::detach
log local0. "XXX LB_FAILED for ip [IP::client_addr]"
...
}
And here is the logged output (IP changed to protect privacy - it's an external static IP address). The events are 4 seconds apart.
Rule /redirect : XXX core URL found, sending to my_pool *.*.*.71
Rule /maintenance-page : XXX LB_FAILED for ip *.*.*.71
Now, my_pool contains two machines and they're both definitely up. Nothing else is triggering LB_FAILED, only this particular client. Connections from *.72, which is physically next to the *.71 one and with the same ISP work. Changing client machines, browsers, etc. does not work, and the same clients can connect from other IP addresses.
Does anyone know why an LB_FAILED might be triggered in this case? The address is not blacklisted.
Thanks in advance