This entries:

Same request: T292

Definitely miss behavior is generated by the new interface MGT that was created when assigning MGT vrf to eth0.

Bug still present in 1.3.0-epa2 version.

I have done a lab using 1.3.0-epa2 version, and got no troubles.
Have configured 2 IP addresses on interface, and configured zone based firewall. I'm able to ping everywhere, using both IP addresses of that interface.
Also I checked that counters on firewall rules got increased while pinging from any of both IP addresses.

Same procedure on 1.2.8, when adding new settings to master, it all remains as master.
Log after commit

Well.. Actually when only one dhcp-server shared network is defined, service can be disbable with global disable for dhcp-server

So, in a very very simple config like this, user is not able to enable and disable dhcp-server with just a simple command?

Error still present in 1.3.0-epa2 version:

Other test.
In same lab, pool for LAN subnet was modified:

More tests where done on version 1.3.0-epa1.
First, dhcp_server.py was modified as indicated.

Resolved using

Tests were done using 1.2.8 and 1.3.0-rc6 version.
Same inconvenient present as described.
Problems is that a default route is added in table 220

Maybe, but if the effort is made in order to be able to configure log level, it would be good that it can be set in different levels.
I'm thinking in a mix scenario, where majority of rules may log with info/debug level (for example default accept rules), while other rules may need a warning/error level (some drop rules).

Bug still present.

At least on simple tests, this modification seems to work.
I have tested on version 1.3, first configuring dhcp server for both addresses: for the one that was defined first, and one for the las IP address defined (of course, one instance of dhcp-server running at a time).
If this needs more testes, let me know what I can do for you.