Experimental Physics and
Industrial Control System

"Al Honey" <[email protected]> · Thu, 18 Nov 2010 08:15:32 -1000

FYI 

The RPC timeout can be controlled with clnt_control (I typically set it
to 3 seconds). I think the default it 22 seconds.

Al

-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Burkhard Kolb
Sent: Wednesday, November 17, 2010 10:02 PM
To: Andrew Johnson
Cc: [email protected]
Subject: Re: How to shutdown/restart a soft-ioc from a (watchdog) record

Am 17.11.2010 um 17:30 schrieb Andrew Johnson:

> Hi Burkhard,
> 
> On Wednesday 17 November 2010 05:03:06 Burkhard Kolb wrote:
>> How do I properly exit a soft-ioc when a record detects some
malfunction?
>> I tried epicsExit(), but that does not work from the subRecord which
I
>> trigger form a calcout record.
> 
> What does the epicsExit() call do when you use it?  That is how you're
meant 
> to shut down the IOC, but there may be drivers running which haven't
properly 
> registered an epicsAtExit() routine to stop their background threads,
which 
> can result in the process not stopping properly.  What extra code are
you 
> including in the IOC?

The extra code is a library which uses RPC (sun...) to communicate. This
is also the
component which apparently hangs until there is a "RPC Timeout" which
actually takes
pretty long.
When the epicsExit call is executed the IOC hangs, the clients report
after some
time the connection loss, but procServ does not restart the IOC because
it's not completely
exited yet.
Burkhard

> 
> - Andrew
> -- 
> If a man is offered a fact which goes against his instincts, he will
> scrutinize it closely, and unless the evidence is overwhelming, he
will
> refuse to believe it.  If, on the other hand, he is offered something
> which affords a reason for acting in accordance to his instincts, he
> will accept it even on the slightest evidence.  -- Bertrand Russell
> 

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System