Experimental Physics and
Industrial Control System

Andrew Johnson <[email protected]> · Fri, 22 Jan 2010 11:22:26 -0600

Hi Caizhen,

On Friday 22 January 2010 10:43:07 Caizhen Liu wrote:
> When I use camonitor to check a PV, I get the following error message:
>
> [dt100@tsdaq1 ~]$ camonitor ACQ196_310_UPTIME
> ACQ196_310_UPTIME              *** Not connected (PV not found)
>
> I have checked the IOC server is running properly.

What is the broadcast IP address as reported by the IOC server machine, and 
does it match the broadcast address that the CA client machine is using?  If 
not you'll need to set EPICS_CA_ADDR_LIST to the IOC's broadcast address for 
the IOC to see your search requests.  The last line of the output from the IOC 
command 'casr 2' should give you the addresses the IOC is listening on.

If the broadcast addresses match, does the IOC server machine run a firewall 
at all, and if so have you opened up the ports 5064 and 5065 to external 
connections?

> Then I try to use
> caSnooper to trouble-shoot. Here is the result:
>
> ---------------------------------------------------------------------------
> [dt100@tsdaq1 linux-x86]$ ./caSnooper -t30 -p4 -c0
> Starting CaSnooper 2.1.2.1 (8-10-2007) at Jan 22 08:27:18
> EPICS 3.14.11
> Individual Name is CaSnoop.test
> Internal PV names are not being published
>
> CaSnooper terminating after 30.01 seconds [0.50 minutes]
>   Data collected for 30.01 seconds [0.50 minutes]
>
> Jan 22 08:27:48:
> There were 20 requests to check for PV existence for 2 different PVs.
>   Max(Hz):   0.33
>   Mean(Hz):  0.33
>   StDev(Hz): 0.33
>
> PVs with top 2 requests:
>    1 192.168.1.1:44539              ACQ196_310_UPTIME                 0.33
>    2 tsdaq1.gat.com:44539           ACQ196_310_UPTIME                 0.33
>
> Connection status for top 2 PVs after 10.00 sec:
>
>
> Questions:
>
> 1)      The IP address 192.168.1.1 and tsdaq1.gat.com of the two requests
> are the address of the client computer. The computer has two Ethernet port.
> Are the ip addresses all right?

That output from caSnooper looks OK to me.

> 2)      Should there be a valid request with the default port number 5064
> instead of the big number 44539?

No, caSnooper is reporting the ports that the requests are coming *from*, not 
where they are going *to*.  There are two port numbers associated with any IP 
packet, the source port and the destination port, and it is pretty rare that 
the two would be the same.  If the client library had to use a *source* port 
of 5064 for all name searches it would not be possible to run more than one 
client on the same network interface, or we'd have to have something like a CA 
repeater to actually send them out; thankfully we don't.

What caSnooper is telling you is that your client appears to be correctly 
sending out PV search requests, but that the IOC is not responding to them for 
some reason.  Check the broadcast addresses as I mentioned above.

- Andrew
-- 
The best FOSS code is written to be read by other humans -- Harald Welte

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System