Experimental Physics and
Industrial Control System

Andrew Johnson <[email protected]> · Thu, 20 Jun 2013 17:42:15 -0500

Hi Lewis,

On 2013-06-20 J. Lewis Muir wrote:
> 
> Right, I understand that.  But I'm (not very clearly) asking about the
> case of writing to a file from an IOC using errlog and not depending on
> another host.  Are you saying one can run iocLogServer on the same host
> as the iocLogClient?  Will this work on vxWorks too?  I hadn't thought
> of doing that.  This would be a new way (to me anyway) of writing to a
> file from one IOC using errlog without depending on another host.

You could definitely do that on any workstation OS.  The current 
iocLogServer.c source code wasn't designed for VxWorks because it has a 
main() function, but there's nothing that inherently prevents you from 
running a similar server on VxWorks.  I would probably want to be sure the 
log file was on an NFS-mounted drive and that the server's task priority is 
sufficiently low to not interfere with the IOC threads, but I don't see any 
other problems.

It appears that feller may not allow you to write errlog messages that don't 
have its special header in them.  If you just want to run a local version of 
the iocLogServer it would be more efficient to write another errlog listener 
program that does the file-writing directly; funneling the messages through a 
TCP socket to a separate local server process is going to waste time 
switching thread contexts which is probably not necessary.

> OK, so here's where I could benefit from using feller: If my goal was to
> write to a file using errlog from an IOC and *not* run iocLogServer at
> all (i.e. not on the same host nor any other host), I could use feller
> to do this.  I think I've got it.

I haven't looked at Benjamin's code so I don't know for sure if that's what 
he has implemented, but we usually think enough alike that I suspect that's 
what feller does.

- Andrew
-- 
Advertising may be described as the science of arresting the human
intelligence long enough to get money from it. -- Stephen Leacock

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System