We have two uplinks (call them A and B) in load-balancing mode, each with 8 static IPs (call them A.0 thru A.7 and B.0 ... B.7).
We've been encountering some weird effects / bugs when trying to migrate some outgoing traffic (SMTP, for example) from egressing on A.2 to B.2, and I now believe this to be because the Masquerading Rules were accidentally conflicting with the Multipath Rules.
This leads me wonder what UTM does when you specify (in the Multipath Rules) that traffic from the SMTP server box goes out on interface B, but tell it (in the masquerading rules) to use a source address from interface A (e.g. A.2).
Obviously, if the configuration is fixed, this shouldn't be an issue in normal operation, but I'm asking because I suppose this issue would break any failover setup, wouldn't it?
I've looked for an explanation of how these two things interrelate in the documentation and in the forums, but can't seem to find anything (I may be blind of course). Is there a writeup anywhere about it? Or does anyone know how it works?
This thread was automatically locked due to age.