[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

19990616: mcidas-os2 & nfs at stc

>From: alan anderson <address@hidden>
>Organization: St. Cloud State
>Keywords: 199906161836.MAA27208 NFS

Alan,

>My student assistant has found that the slow data transfer rates 
>that we experienced within our local system appear to have been
>solved.  We assume you people solved the problem.

I wish we could take credit for this one, but we can't.

>I have not 
>checked myself since I just returned today and I notice that we
>stopped getting mcidas data from our feedsite at unl as of late
>yesterday.  I have sent an email to Clint Rowe, but no change
>as of this time.  

OK.

>What did you find?

We found that if we ignored a problem long enough it would clear itself
up.  Seriously though, did your student happen to reboot the OS/2 machine
and/or the Solaris x86 box (waldo)?

>We will want to change our other os2 machines
>over to waldo as well so will need to fix them as well.

Sounds good.  Just to keep you uptodate with fiddling that I have been
doing on your machine: I uploaded a new Unidata-Wisconsin datastream
image decoder to waldo.  The reason was that the version that I had put
there previously had a problem that manifested itself in horizontal
streaks across images and "speckles" in the outer space portion of
visible images.  The problem appears to be confined to code compiled
with gcc/f2c on Solaris x86.  No other platform including Solaris x86
using Sun compilers showed this problem with 'lwtoa3'.  I spent close
to 4 days trying to find out exactly what was wrong before finally
giving up and making a code mod to a McIDAS routine that got around the
problem.  I am still scratching my head over this...

Tom

>From address@hidden  Fri Jun 18 08:43:58 1999

I had noticed the streaks, just figured it was holes in the data; as to the
speckles, hey, I wondered if the sensors were picking up starlight.

My assistant has been gone this week, so I am not sure what he did as to 
rebooting.  I do not think waldo was rebooted.  

The next steps I plan are

1. Re try the connection of our other os2 machines to waldo.  We have
   put them back onto hobbes in light of the slow data transfer.

2. Begin the process of converting to solarisx86  and mcidasx on the pc's
   that now  run os2.  I am assuming we can let them get their data from 
   waldo just as they get it from hobbes at present.

All comments and suggestions are appreciated.

Alan