Hi
Pierrick,
First, I admit some details concerning my philosophy on error messages; My perspective is that users aren’t dumb and
if they are provided with low level details concerning failures this can sometimes be quite useful when determining the causes of problems.
Ø
When I run my
iocs, I get a lot of CA beacon errors when I run my
iocs.
Ø
../online_notify.c: CA beacon (send to "148.79.117.13:6724") error was "Connection refused"
In the CA server a connected mode UDP socket is used to send the beacon messages. This can provide some additional diagnostics
when there are routing problems with the UDP messages; in essence, any Internet ICMP diagnostics messages that might occur can sometimes precipitate into useful low-level error number information returned from the connected mode UDP socket send call.
This “Connection Refused” message is in fact the direct to string translation of the error number that the socket send
call has returned.
With the CA beacon UDP messages there can be unicast (single receiver) and broadcast (multiple receiver) destination
addresses. With the beacon messages we hope that the destination side is the CA (Beacon) Repeater daemon on behalf of the CA Client Library. In the diagnostic message you are receiving we have a
unicast receiver side address. In this situation the software is informing us, indirectly via ICMP, that there isn’t any UDP server awaiting a unicast message at
148.79.117.13:6724.
The CA Server’s list of destination addresses for beacons is typically auto-configured. The server typically sends to
port 5065 (the CA Repeater’s port) at the broadcast addresses destination of each of its locally attached network interfaces. One can add some additional destination addresses to this list using the EPICS_CA_ADDR_LIST in the legacy 3.14 IOC’s server, and using
the EPICS_CA_BEACON_ADDR_LIST environment variable (backwards compatible defaulting to the original EPICS_CA_ADDR_LIST variable) in the PCAS server and its decedents.
So the bottom line question is this one; why is a CA server sending a beacon message to
148.79.117.13:6724. This is somewhat unusual situation because this
port 6724 is probably a CA server destination port and not a CA Repeater Daemon destination port.
Possibly this has happened because of non-default configuration in the server’s EPICS_CAS_BEACON_PORT variable (defaulting
to the legacy EPICS_CA_REPEATER_PORT variable), or perhaps specific port numbers are supplied along with the IP addresses in the EPICS_CAS_BEACON_ADDR_LIST variable (only for the PCAS server which does not run in the IOC) or the EPICS_CA_ADDR_LIST variable
(used by the legacy server in the IOC).
Currently, I am unable to draw specific conclusions concerning the 148.79.117.13 address because I don’t yet know what
might be running on this host (the server or the client or both).
Jeff
Hi all,
I have a few soft iocs I'm running on a linux box running SL 2.6.18... My epics
version is 3.14.11. I have a similar configuration for another control system elsewhere.
Because I have multiple servers on a single processor, I generally assign them different
ports (initially set up by a former colleague). The ports I'm using are 6721, 6722, 6723,
and 6724. The firewall is not set up on this machine, so I don't know if there is something
else that I should be doing in the configuration.
When I run my iocs, I get a lot of CA beacon errors when I run my iocs.
../online_notify.c: CA beacon (send to "148.79.117.13:6724") error was "Connection refused"
../online_notify.c: CA beacon (send to "148.79.117.13:6722") error was "Connection refused"
../online_notify.c: CA beacon (send to "148.79.117.13:6724") error was "Connection refused"
../online_notify.c: CA beacon (send to "148.79.117.13:6722") error was "Connection refused"
../online_notify.c: CA beacon (send to "148.79.117.13:6724") error was "Connection refused"
I've spent a lot of time reading the CA manual and looked through techtalk and haven't
found a definitive solution to fixing this. I'm under the impression that I need to set the
beacon environment variable, but to what?
Thanks,
Pierrick
--
"Whether you think you can or think you can't, either way, you are correct" -- Henry Ford
_______________________________________________________________
Pierrick Hanlet
IIT/Fermilab
+1-630-840-5555 (FNAL)
+1-312-567-5745 (IIT)
+1-630-697-8758/+44-79-48-860-197 (US/UK mobile)