Experimental Physics and
Industrial Control System

Marty Kraimer <[email protected]> · Wed, 21 Jan 1998 10:14:40 -0600

Ralph Lange wrote:
> 
> A colleague from the diagnostics group encounters the following problem on
> one of his IOCs:
> 
> The IOC is the central point of some data aquisition process. I.e., the
> database is connected to 16 other IOCs with one link from and one link to
> each IOC.
> 
> There is a sequencer state machine supposed to be running on the IOC that
> sets up one "assign" connection to an external PV (not on this IOC), three
> assigns to PV on the IOC itself.
> 
> Without loading the state machine, the IOC runs without error.
> 
> When the state machine is loaded, approx. ten seconds later (obviously
> while the CA clients builds the connections) the IOC starts printing error
> messages of the form "CAC: ... Unable to allocate a new socket ...". The
> failing CA connections correspond to CA links contained in the database,
> not the state machine.

we modified our board support package to allow more file descriptors.
I think we made the limit 150.

Marty Kraimer

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System