Re: Remote IOC reboot [was: EPICS for implementation of Remote Procedure Call]

[Ralph Lange <Ralph.Lange@helmholtz-berlin.de>]

Hi Emmanuel,

as Stephanie points out, a simple sub record allows for CA based reset 
on most IOC platforms. (No special device support necessary.)
Remember, though, there are cases when this will or might not work: IOC 
hangs (so the IOC doesn't work), IOC gets CPU saturated (so that CA 
doesn't work), IOC gets memory saturated (so that CA doesn't allow new 
connections). In those cases, you will probably have to get to the 
console of the IOC and do whatever necessary (e.g. type ^X). Workaround: 
it is relatively easy to create a shell script that connects to the 
console and echoes a ^X to it, so you can still create a GUI button that 
calls the shell script that connects to the console and reboots the IOC.
Less elegant than CA to a sub record, still more efficient than walking 
over.
Also, for VME based systems, there are situations that need a VME reset 
rather than just a reboot of the CPU.
Some VME crates (e.g. those by Wiener) are remotely controllable, so 
that the whole VME crate can be switched on/off or reset remotely (plus 
a bucket load of other functionality).
For other (aka less expensive) crates some labs have developed VME cards 
that can be remotely accessed and are able to pull the VME reset line 
(plus more functionality). That is not free: A separate card always uses 
an extra slot.
Cheers,
Ralph


On 15.04.2009 20:42 Allison, Stephanie wrote:
> Hi Emmanuel,
>
> There are some device support packages that provide for rebooting/restarting an IOC via record.  For instance, devIocStats:
> http://www.slac.stanford.edu/comp/unix/package/epics/site/devIocStats/
> provides a sub record that calls epicsExit for soft iocs, reboot for vxWorks iocs, and bsp_reset/rtemsReboot for RTEMS iocs.  
>
> Stephanie Allison
>
>   
> > -----Original Message-----
> > From: tech-talk-bounces@aps.anl.gov 
> > [mailto:tech-talk-bounces@aps.anl.gov] On Behalf Of Emmanuel Mayssat
> > Sent: Wednesday, April 15, 2009 11:26 AM
> > To: banerian@u.washington.edu
> > Cc: Mahajan Kirti; tech-talk@aps.anl.gov
> > Subject: Remote IOC reboot Was(EPICS for implementation of 
> > Remote Procedurecall)
> > S,
> >
> > Could you describe your interface (from a user/operator 
> > perspective) to
> > your IOC reboot interface? I may actually be interested with that.
> >
> > In short I want to know:
> > Can you reboot/control a remote IOC with the push of a GUI button?
> >
> > --
> > Emmanuel