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

20010228: ADDE serving of sounding data on RedHat 6.2,7.0 (cont.)



>From: Unidata Support <address@hidden>
>Organization: UNCA
>Keywords: 200102052335.f15NZkX27276 McIDAS-XCD

Leigh,

I just wanted to touch base with you regarding the problem of serving
sounding data from a remote ADDE server hosted on RedHat 6.2/7.0
Linux.

I have periodically logged onto storm2 and run different tests.  At one
time I was convinced that I had pushed off the failure of the service
to the point that a remote client would get enough data to do a display
(sounding, hodograph, upper air listing).  To get to this point, I had
made a modification in the mcserv module.

In a follow-up test this morning, I tried reverting back to the 7.7x
version of mcserv to see if the failure would return to "normal".  To
my suprise, I was and am still able to access sounding data off of your
machine to the point where plots get made.  I still get a 'pipe error
Connection reset by peer' error indication in my session when using
compressed transfers, but the soundings, hodographs and upper air lists
have what looks to be a full set of data (or if it is not complete, it
is damn close).

A 'uname -a' on storm2 shows that you are once again running the 2.2.18
kernel.  The question I now have is if you have done anything else to
storm2 that might possibly affect the ADDE services (affect how long a
socket can/will be kept alive)? If not, the serving mystery continues
to deepen!

Just to catch you up on all of the things I have tried, I can report
that the only systems that I have seen the failure on are:

RedHat 6.2 and 7.0 Linux  (I don't have access to 6.0 or 6.1, so I can't
                           say whether or not I would see the failure there)

I have verified successful serving off of:

RedHat 5.2 Linux
FreeBSD 4.2
DEC OSF/1 4.0 E,F
HP HP-UX 11.00
IBM AIX 4.3
SGI IRIX/IRIX64 6.2.5
Sun Solais x86 2.[78]
Sun Solaris SPARC 2.[678]

My next move is to try some tests on Debian Linux to see if there are
any problems there.  After that, I will be bugging RedHat and/or Linux
mail groups (sigh).

Tom

>From address@hidden Wed Feb 28 12:33:06 2001
>Subject: Re: 20010228: ADDE serving of sounding data on RedHat 6.2,7.0 (cont.) 

Tom,

Sheesh... I am baffled. I, too, can get soundings come up with
or without compression... with the exception of one time with an
uncompressed transfer where it gave me the usual error.

I haven't gone back to 2.4.0 since that reboot a few weeks ago (lastlog
shows Feb 6) when you thought that it might have something to do with
the new kernel.  The machine was rebooted 8 days ago, dunno why,
probably was acting flaky so someone rebooted it (someone logged in as
unca-mcidas, I think).

So I guess the mystery deepens. I haven't upgraded any software for a
while, either, I think the last time was Jan 17. The fact that I was
able to get one instance of it not working suggests that the problem
still exists, but I sure can't explain the irregularity. Basically I
haven't done squat with storm2 in the past few weeks. Students have been
regularly using it for mcidas under the unca-mcidas account and for
simple unix tasks... but that's it.

Leigh
http://orf.cx