Re: caget: Identical process variable names on multiple servers

[Michael Davidsaver via Tech-talk <tech-talk@aps.anl.gov>]

On 11/11/25 12:41 PM, Smith, Martin via Tech-talk wrote:

As far as if you are getting strange IP addresses I'm not sure what that is about.

Is this search reply coming from a nameserver?

The extension which allows for CA nameservers did so by changing the meaning of the SID field in the search reply message.  Previously, this field contained a server assigned channel identifier which, historically servers start counting from 1.  So seeing a low numbered value makes me wonder if something is triggering some compatibility code somewhere.

https ://docs.epics-controls.org/en/latest/internal/ca_protocol.html#id2

In the IOC (RSRV) this code path has been stubbed out.  Beginning with Base 3.16.0.1, an IOC will always respond with SID==0xffffffff .

https ://github.com/epics-base/epics-base/blob/b308be3273fef5dc3b3e6af21bde91e409454d8f/modules/database/src/ioc/rsrv/camessage.c#L2193-L2201

A corresponding change was made to PCAS at the same time, so I think a ca-gateway would also reply with 0xffffffff.  However, a nameserver built on PCAS could reply with any IP.

Before 3.16.0.1, an IOC (or gateway?) receiving a search with a CA minor version less than 3 would reply with the SID number.