Fix breakage for smartswitch usecase

[venkit-nexthop]

What I did
Make fpmsyncd use the mgmt vrf to bind to zmq when configured.

Why I did it
Refer to sonic-net/sonic-buildimage#19638 The previous PR (#3979) inadvertantly caused the above issue.
This commit fixes that by connecting to zmq from fpmsyncd using the vrf as well (if one is configured)

How I verified it
By ensuring that when mgmt vrf is configured, FRR routes are programmed correctly via orchagent.

I also ran test_gnmi_zmq.py with the following diff and it ran successfully.

--- a/tests/zmq/test_gnmi_zmq.py
+++ b/tests/zmq/test_gnmi_zmq.py
@@ -28,6 +28,7 @@ def save_reload_config(duthost):
     orchagent_pid = get_pid(duthost, "orchagent")
     telemetry_pid = get_pid(duthost, "telemetry")

+    duthost.shell("sudo config vrf add mgmt", module_ignore_errors=True)
     result = duthost.shell("sudo config save -y", module_ignore_errors=True)
     logger.debug("Save config: {}".format(result))
     result = duthost.shell("sudo config reload -y -f", module_ignore_errors=True)

Details if related

[mssonicbld]

/azp run

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).