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

20030124: LDM install problem on Solaris 8 at Georgia Tech

>From:  "Jim St. John" <address@hidden>
>Organization:  Georgia Tech
>Keywords:  200301241750.h0OHoEx05986 LDM install

Hi Jim,

>I'm trying to install ldm 5.2 on a new Sun Sparc Server running SunOS 8.
>I've installed it twice double checking every step of the preinstall
>checklist and installation checklist.

Are you doing a binary or source installation?  If you are doing a
binary installation, and continue to run into problems, I would
recommend building from source.

>When I run ldmadmin start nothing
>happens until I get an ldm not started or registered message. Nothing is
>written to the ldmd.log.

I am assuming by this comment that ~ldm/logs/ldmd.log exists and
is writable.  Is this true?

A common startup problem with the LDM is some sort of typo in the
~ldm/etc/ldmd.conf file:  those problems will cause 'ldmadmin start'
invocation to hang, and no LDM processes will get started.  Diagnosing
this problem relies on being able to log to ~ldm/logs/ldmd.log, which,
in turn, depends on the file existing, being writable, AND syslog.conf
being setup correctly.

Did you setup syslog.conf as per LDM installation instructions and
remember to send a HUP to the process ID of syslogd?

If the answer is yes, then you can test to make sure that syslogd will
log local0.debug to ~ldm/logs/ldmd.log by using the system utility
'logger'.  Here is a simple 'logger' invocation, run as 'ldm', that
will test to see if syslogd is setup to log to your LDM log file:

% logger -p local0.debug "Test of syslogd output to LDM log file"

If you do not see the message listed at the end of your LDM
log file, then one or more of four things may be the cause:

o ~ldm/logs/ldmd.log is not writable
o /etc/syslog.conf is not setup correctly
o you did not send a HUP to the process ID of syslogd
o syslogd is hung and needs to be stopped and restarted by 'root'

>I have run ldmadmin pqactcheck and it says the
>pqact.conf file is syntactically correct.

If no LDM processes are started (check this by running 'ps -eaf | grep ldmd'),
then the problem is usually in ~ldm/etc/ldmd.conf.  Once syslogd
is logging correctly, we should be able to see error messages in
the LDM log file that will pinpoint errors in ldmd.conf.

>I have double checked the write
>permissions on the data directories.
>
>Has anyone encountered these symptoms?  What should I be checking.

Solaris 8 is one of the most commonly used platforms for the Unidata
LDM.  One of our main IDD nodes, motherlode.ucar.edu, is running
Solaris 8, and we have not experienced the LDM
installation/configuration problems you are seeing.

Please keep us informed as to your progress.

Tom
--
+-----------------------------------------------------------------------------+
* Tom Yoksas                                             UCAR Unidata Program *
* (303) 497-8642 (last resort)                                  P.O. Box 3000 *
* address@hidden                                   Boulder, CO 80307 *
* Unidata WWW Service                             http://www.unidata.ucar.edu/*
+-----------------------------------------------------------------------------+