EPICS Controls Argonne National Laboratory

Experimental Physics and
Industrial Control System

1994  1995  1996  1997  1998  1999  2000  2001  2002  <20032004  2005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024  Index 1994  1995  1996  1997  1998  1999  2000  2001  2002  <20032004  2005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024 
<== Date ==> <== Thread ==>

Subject: RE: PowerPC woes
From: "Woods, PD (Paul)" <[email protected]>
To: "Epics Tech Talk (E-mail)" <[email protected]>
Date: Fri, 17 Jan 2003 14:16:45 -0000
Hi,

Some months ago this thread allowed us to solve this same problem at our
site. However due to a hardware fault our Unix boot server was lost. One new
harddisk, OS upgrade (to Solaris 9) and data recovery from tape later our
very long IOC boot problem has now returned! We have also forgotten which
startup script we put this line in, anyone any suggestions?

Thanks

Paul Woods
Diamond Project UK

-----Original Message-----
From: Rozelle Wright [mailto:[email protected]]
Sent: 21 November 2001 20:58
To: [email protected]
Cc: [email protected]
Subject: Re: Fwd: RE: PowerPC woes


Now that I'm retired, I don't check my e-mail as often as I used to, so Nick
has probably solved his problem, but since there were so many different
solutions, I thought I would tell you what I know.  This problem has the
same
symptoms as we have had here at Los Alamos with slow booting of the baja60
IOCs
from Solaris machines.  We removed the passwords from those IOCs inside the
firewall, so that they would use tftp to boot and that solved the problem
there.  For those baja60s outside the firewall where allowing tftp on the
Solaris machines was not an option we arranged for them to boot from Suns
running Solaris 7 or later and made a change to the ndd parameter on the
Suns.
Our particular magic line in a startup files for the suns is

/usr/sbin/ndd -set  /dev/tcp tcp_time_wait_interval 5000.

(Note:  /usr/sbin/ndd doesn't have tcp_time_wait_interval as a valid
parameter
on Solaris 6 or earlier  Solaris.)

We decided that the problem was caused by an attempt to use the same TCP
port
too soon.

In order to solve this problem using either of these two solutions you need
to
be or get help from someone who is super user on the Sun from which the iocs
are booting.

Both of the suggestions above were suggested in previous e-mails, but
referred
to different types of  IOCs.  I suspect the problem is a generic problem
with
vxWorks.

Rozelle Wright
[email protected]




"Dale L. Brewe" wrote:

> >Another possible cause of this problem could be that
> >a boot was interrupted while in progress. If the IOC
> >comes up and tries to reuse the same TCP port the
> >Solaris server sometimes waits a very long time before
> >it times out the old TCP circuit and begins communicating
> >with the IOC which is presenting TCP sequence numbers
> >that are invalid in the Solaris FTP server's perspective.
> >
>
> I haven't been following this too closely, but this suggestion reminds me
> of the problems we started  having after we upgraded to solaris 2.7 from
> 2.5.1. The problem went away after issuing the command "/usr/sbin/ndd -set
> /dev/tcp tcp_time_wait_interval 10000" . I think the default was 240000.
> This doesn't survive a reboot, so I put this is a startup script. I
believe
> notes on this are somewhere in the EPICS listserver archives.
> dale Brewe
>
> > >
> > >
> > > We had a similar problem when booting our MV2304's where it would
> > > take an extremely long time to boot.  After messing with the ethernet
> > > settings we finally settled on changing from using rsh to tftp when
> > > booting our crates:
> > >
> > >       ftp password = blank
> > >       flags           = 0x88
> > >
> > > Of course this means you need to have your host setup to support tftp.
> > > In our case our hosts are Solaris workstations.
> > >
> > > Kevin Tsubota
> > >
> > >
> > > > -----Original Message-----
> > > > From: Nick Rees [mailto:[email protected]]
> > > > Sent: Friday, November 16, 2001 10:39 AM
> > > > To: Mark Rivers
> > > > Cc: '[email protected] '
> > > > Subject: RE: PowerPC woes
> > > >
> > > >
> > > > On Fri, 16 Nov 2001, Mark Rivers wrote:
> > > >
> > > > > I have a different suggestion from Andrew's.  I have seen a
> > > > very similar
> > > > > problem to the on Nick describes on the MVME2700.  The
> > > > problem is actually
> > > > > with the Ethernet 10/100 autonegotiation.  If I plug the
> > > > Ethernet into a
> > > > > 10/100 Hub it works fine.  If I plug it into certain types
> > > > of Ethernet
> > > > > switches (e.g. Cisco 2924 I think) it takes several
> > > > minutes, and it does it
> > > > > twice as Nick observed. Try using a different physical
> > > > device and see if it
> > > > > improves.
> > > >
> > > > I have tried four configurations:
> > > >
> > > > 1. An autoconfiguring 10/100 switch (Allied Telesyn),
> > > > 2. An autoconfiguring 10/100 hub (SMC)
> > > > 3. A fixed 10 base T switch (Allied Telesyn)
> > > > 4. A fixed 100 base TX hub (Allied Telesyn).
> > > >
> > > > All with no result. Sob. Sob.
> > > >
> > > > I have also tried changing the address translation stuff as
> > > > per Andrew's
> > > > suggestions, but to no avail. However, I may have got it
> > > > wrong, so I will
> > > > continue to think about it.
> > > >
> > > > Cheers,
> > > >
> > > > --
> > > > Nick Rees
> > > >
> > > > Joint Astronomy Centre               Ph:       +1 (808) 961-3756
> > > > 660 N. Aohoku Place                  Fax:      +1 (808) 961-6516
> > > > Hilo, HI.  96720                     Internet:
[email protected]
> > > >
>
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>   Dale L. Brewe     PNC-CAT/U. of Washington
>   ph: 630 252 0582         fax: 630 252 0580
>   Mail:
>   Argonne National Laboratory
>   Bldg 435E sector 20
>   9700 S. Cass Ave.
>   Argonne, IL 60439
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Replies:
RE: PowerPC woes Andy Foster

Navigate by Date:
Prev: Re: EPICS exmaple application problem M.C.Shao
Next: GSOctal 232 reboot problem Woods, PD (Paul)
Index: 1994  1995  1996  1997  1998  1999  2000  2001  2002  <20032004  2005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024 
Navigate by Thread:
Prev: Fw: Re:RE: EPICS exmaple application problem M.C.Shao
Next: RE: PowerPC woes Andy Foster
Index: 1994  1995  1996  1997  1998  1999  2000  2001  2002  <20032004  2005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024 
ANJ, 10 Aug 2010 Valid HTML 4.01! · Home · News · About · Base · Modules · Extensions · Distributions · Download ·
· Search · EPICS V4 · IRMIS · Talk · Bugs · Documents · Links · Licensing ·