20030610: trouble with McIDAS ADDE and RH 8.0

>From: Robert Mullenax <address@hidden>
>Organization: NMSU/NSBF
>Keywords: 200306102151.h5ALprLd028481 McIDAS ADDE RedHat 8.0 Linux

Hi Robert,

>We got some new Dells here at NSBF.  I had intended to put Solaris on
>them, but it won't install, despite installing on a server that is very
>similar.  So I used RH 8.0.

OK.  Several other sites are running under RedHat 8.0 (and 9.0!)
so there are not endemic problems running McIDAS on that (those) OS.

>I have McIDAS-X built and installed.but I can't get ADDE to work.  I
>can't communicate with any site.  The machine I am using (setting it
>up) is not behind a firewall right now.
>
>If I do a DSINFO X RTIMAGES (where RTIMAGES is pointed at
>wxmcidas.nsbf.nasa.gov) I get:

>[mcidas@newpsn ~]$ dsinfo.k X RTIMAGES DEV=CCC
>dsinfo.k* ***** REQ. CFILE-->ALA.RTIMAGES<--
>dsinfo.k* Read of server has failed. Wanted  4
>dsinfo.k*                               Did
>    No Datasets found of Type: IMAGE in Group: RTIMAGES
>dsinfo.k* ***** REQ. CFILE-->ALM.RTIMAGES<--
>dsinfo.k* Read of server has failed. Wanted  4
>dsinfo.k*                               Did
>    No Datasets found of Type: POINT in Group: RTIMAGES
>dsinfo.k* ***** REQ. CFILE-->ALV.RTIMAGES<--
>dsinfo.k* Read of server has failed. Wanted  4
>dsinfo.k*                               Did
>    No Datasets found of Type: VPROF in Group: RTIMAGES
>dsinfo.k* ***** REQ. CFILE-->ALG.RTIMAGES<--
>dsinfo.k* Read of server has failed. Wanted  4
>dsinfo.k*                               Did
>    No Datasets found of Type: GRID  in Group: RTIMAGES
>dsinfo.k* ***** REQ. CFILE-->ALT.RTIMAGES<--
>dsinfo.k* Read of server has failed. Wanted  4
>dsinfo.k*                               Did
>    No Datasets found of Type: TEXT  in Group: RTIMAGES
>dsinfo.k* ***** REQ. CFILE-->ALN.RTIMAGES<--
>dsinfo.k* Read of server has failed. Wanted  4
>dsinfo.k*                               Did
>    No Datasets found of Type: NAV   in Group: RTIMAGES
>DSINFO -- done
>[mcidas@newpsn ~]$

Very strange indeed.  What happens for a locally defined dataset?

>That happens with any site I try..yet I can ssh to those sites and I
>successfully got an LDM feed from wxmcidas.

Got it.

>I looked in the archives and saw a message from June 3 about an issue
>on RH 7.3.  I also went in and removed RECORD from the three mc* files
>in /etc/xinetd.d.

Right.  This was a strange situation that popped up at Millersville.
I have not experienced the same error here at Unidata on either RH
7.3 or 8.9.

>When I do xinetd -d I get this strange bind error..
>towards the bottom..and then that CRITICAL error at the end.. BTW I
>installed all the latest patches on this box including an xinetd
>patch.
>
>
>03/6/10@21:48:40: ERROR: 1376 {activate_normal} bind failed (Address already i
> n
>use (errno = 98)). service = mccompress

My first inclination is to question whether or not /etc/services has more
than one definition for ports 112, 500, and 503.  Port 112 is registered
as 'mcidas' (by us, no less), and ports 500 and 503 can be registered
by other services.  Here is what I would do:

1) edit /etc/services and comment out the duplicate entries for any/all
   of ports 112, 500, and 503

2) make sure that there is not a non-McIDAS application named mcserv
   that is already running 

3) make sure that /etc/xinetd.conf contains a reference to the /etc/xinetd.d
   directory.  Here is what /etc/xinetd.conf looks like on our RH 7.3
   system:

#
# Simple configuration file for xinetd
#
# Some defaults, and include /etc/xinetd.d/

defaults
{
        instances               = 60
        log_type                = SYSLOG authpriv
        log_on_success          = HOST PID
        log_on_failure          = HOST
        cps                     = 25 30
}

includedir /etc/xinetd.d

One quick point: the mcserv, mccompress, and mcidasz files in
/etc/xinetd.d have nothing to do with McIDAS running on the machine
accessing a remote ADDE server.  They are only used by remote systems
accessing the remote ADDE server on the local machine.  Along with that
will be the allows in /etc/hosts.allow: they should only be used for
remote machine access to the ADDE serve on the local machine.

If all else fails, I am more than willing to logon and poke around.

Tom