We run the SystemEDGE in standalone mode on RHEL7 and configure it to bind to a specific IP address using the 'bind_address' in the sysedge.cf file.
We noticed that in addition to the default_port (1691) we see the agent binding to two additional high udp ports and for one of those two (56102), does not honor the bind_address setting:
# ss -ulpn
State Recv-Q Send-Q Local Address:Port Peer Address:Port
UNCONN 0 0 6x.xxx.xxx.xxx:1691 *:* users:(("sysedge",pid=13988,fd=8))
UNCONN 0 0 6x.xxx.xxx.xxx:32932 *:* users:(("sysedge",pid=13988,fd=7))
UNCONN 0 0 *:56102 *:*
I couldn't find any reference to those two ports in the manual and they change upon restart.
Can you please tell me what the agent uses these for and if it's possible to have both bind to the specified address?
Release : SystemEDGE 5.9
You can restart the SystemEDGE agent and check the sysedge.log as long as you see the print that the agent is binding to the correct address this is all you need to worry about.
0000028 2021-09-08 14:52:08.18 [I]-b44- se/sysedge_config.c : parse_config_pass0(): Using '10.7x.xx.xxx' as bind address
You can bind SystemEDGE to a specific IP address and the agent will use that IP for incoming SNMP requests on port 161 or 1691 but the other ports are ephemeral ports that are automatically negotiated by the operating system those do not need to be managed.