20050901: [SSEC-TC-req #10653] syslogd on amrc.ssec.wisc.edu (cont.)

>From: Matthew Lazzara <address@hidden>
>Organization: SSEC/AMRC
>Keywords: 200509011411.j81EB6jo005065 LDM

Hi Matthew,

>Thank you every so much for this!!  Interesting, my env didn't have  
>any of those environmental variables set,

Interesting...

>but going through your  
>steps, seems to have solved the problem.

What C compiler is being used (e.g., Sun's C compiler or GNU C)?

>I wonder if attempting to  
>start ldm before doing the make install_setuids was a bad thing...

The 'make install_setuids' step is needed to set the ownership
and permissions on 'hupsyslog' and 'rpc.ldmd', both of which must
run (initially) as 'root'.

'hupsyslog' sends a signal (HUP) to the syslogd daemon telling it to
close open file descriptors; reread its configuration file
(/etc/syslog.conf); and then reopen files.  This is needed since we use
syslogd to write to the LDM log file, ~ldm/etc/ldmd.conf.

'rpc.ldmd' runs as 'root' just long enough to open port 388.  After the
port is open, it runs as 'ldm'.

>In any case, I have restarted my ldm, and didn't get that killall  
>error, which I had been getting...

I am still suprised by the 'killall' error.  The build on Solaris
should _always_ try to run 'kill -HUP <pid of syslogd>', not
'killall'.  On systems where there is no file containing the process ID
of syslogd, 'killall' is run to send the signal to the process by
name.  Since the process ID is contained in /etc/syslog.pid under all
versions of Solaris, 'hupsyslog' should never try to run 'killall'.  My
previous comment about environment variables being set before
'configure' is run was based on an exchange with another user who had
migrated from an SGI box to Linux.  In that case, there were a number
of environemnt variables set (CFLAGS and/or CPPFLAGS in particular)
that intefering with configure's ability to figure out if the syslogd
PID was available in a file.

>Thank you all again for the help with this!!

No worries.

By the way, you should not be having problems building LDM-6.4.x on
this same system (you made a passing comment about this in your
previous email).  You might try building 6.4.1 in the same way that you
built 6.3.0 to verify that the problems have gone away.

Cheers,

Tom
--
NOTE: All email exchanges with Unidata User Support are recorded in the
Unidata inquiry tracking system and then made publicly available
through the web.  If you do not want to have your interactions made
available in this way, you must let us know in each email you send to us.

>From address@hidden  Thu Sep  1 10:37:30 2005

Hi Tom!!

The C compiler I've got on this Sun is Sun's WorkShop Compilers 5.0 -  
I'm still running Solaris 8 (my poor Ultra 10 is getting on 5 years  
old or so...but replacing this is planned, pending NSF funding...late  
2006 if we're lucky).

I might try 6.4.1 again, but the build didn't work at all when we  
tried. (filed during the "make"/compile step).

Thanks again Tom for the support!

Best Regards,

Matthew