Experimental Physics and
Industrial Control System

Pete Jemian <[email protected]> · Fri, 17 Apr 2015 10:30:06 -0500

<my_own_opinion>

Reboot of the workstation hosting the soft IOC should not, by design, be

provided so easily to an IOC for the reasons previously stated.

What differs between "soft" IOCs and standard IOCs?

On a standard IOC (such as EPICS running on a VxWorks operating system

on a VME board), the IOC software is the only software suite on that

computer.

An EPICS soft IOC is one of potentially many software suites running on

the host computer. Others suites would include user login sessions and

non-EPICS services provided to the computer, LAN, or WAN.

For standard IOCs, reboot of the computer seems natural.

For soft IOCs, reboot of *only* the soft IOC (not the computer) seems

natural.

</my_own_opinion>

Pete

On 4/17/2015 9:32 AM, Shen, Guobao wrote:

Hi Geyang,
The reboot function used in iocStats has been redefined as for example
"#define reboot(x) epicsExit(0)" on any posix system.
You can find full definition in file for example
devIocStats/os/posix/devIocStatsOSD.h
Therefore, the reboot is a alias name of epicsExit() instead of not a
linux OS reboot, and does accept parameter.

I hope this is your question.

Guobao

On 4/17/15 9:20 AM, Silver wrote:

hi, Ralph:

Thanks a lot for your reply.
maybe you misunderstand my question for my poor english. I think
reboot function should restart the whole machine, but why the reboot
function here just kill the IOC process?
In the man help about reboot function, I didn't see any parameters for
just rebooting the process?
------------------------------------------------------------------------
Best wishes
Geyang 2015-04-17

    *From:* Ralph Lange <mailto:[email protected]>
    *Date:* 2015-04-17 16:44
    *To:* EPICS Tech-Talk <mailto:[email protected]>
    *Subject:* Re: iocStats reboot question
    Hi Geyang,

    Under Linux (and the other host-type OSs), an IOC is a single
    multi-threaded process.
    So, naturally, "rebooting" an IOC means restarting the IOC
    process, which is usually done by shutting down or killing the
    current process and letting the wrapper (e.g. procServ) handle the
    restart.

    Very often, many Linux IOCs are run on a single host: 20-30 are
    quite common, I have seen up >300 IOCs on a server-class Solaris
    machine.
    I am sure you understand that allowing to reboot the server from
    any of these IOCs would be a bad idea.

    Also, rebooting a server usually needs root access. IOCs should
    not be run as root.

    Cheers,
    ~Ralph

    On 17/04/2015 09:36, Silver wrote:

    hi,

    in IOC Status and Control (devIocStats), it has an reboot
    function in subroutine record, that is ,in devIocStatsSub.c file,
    it calls reboot function of linux.
    I am wandering why this call just kill the ioc process instead of
    rebooting the machine? as in manual of reboot function
    (http://man7.org/linux/man-pages/man2/reboot.2.html)
    Thanks a lot for your help.

    ------------------------------------------------------------------------
    Best wishes
    Geyang 2015-04-17

--
----------------------------------------------------------
Pete R. Jemian, Ph.D.                <[email protected]>
Beam line Controls and Data Acquisition, Group Leader
Advanced Photon Source,   Argonne National Laboratory
Argonne, IL  60439                   630 - 252 - 3189
-----------------------------------------------------------
   Education is the one thing for which people
      are willing to pay yet not receive.
-----------------------------------------------------------

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System