Installation of DMT rev_2.1.0pre

Patrick J. Sutton   Date: Wed Jul 3 19:18:30 2002 (GMT)

This log file contains the steps used to install DMT rev_2.1.0pre (2002/07/03 version) on a Dell workstation named 'dmt' running Red Hat Linux 7.2.


Preparations

The various packages are or will be installed in the following directories:

All of the utility programs (gcc-2.95.3, libtool 1.4, make 3.79.1, ROOT) are already installed (see bottom for details). ROOT 3.02.07 is still the current production version.

To prepare for the installation of rev_2.1.0 I have also done the following:

Finally, note that /etc/profile.d/dmt.csh points to rev_2.0.0; this will have to be changed after successfully upgrading to 2.1.0.

Installation of rev_2.1.0pre

The following procedure was found to work for installing rev_2.1.0pre:

Preparations for DMT Installation

Notes from the installation of the utilities required by the DMT. These are

Building GCC-2.95.3

Make sure we are using gcc and its libraries. We will need to add to /etc/profile.d/{path,env}.{sh,csh}; for now, add to ~dmt/.cshrc

  set path = (/usr/dmt/bin $path) 
  setenv LD_LIBRARY_PATH /usr/dmt/lib 

LibTool Install

Installing LIGOTOOLS

Will need to add "eval `/usr/dmt/bin/use_ligotools`" to start-up environment (e.g., in /etc/profile.d/env.{sh,csh}). For now, put in ~dmt/.cshrc

ROOT Install

Need to add /usr/dmt/root/bin to path, also /usr/dmt/root/lib to LD_LIBRARY_PATH. This can be done by sourcing $DMTHOME/$DMTVERSION/bin/setup, setupb, and root-setup in a file(s) in /etc/profile.d/ once we have the DMT installed; put in .cshrc for now.

ROOT appears to be working. On to DMT.



Correspondence with J. Zweizig

From jzweizig@ligo.caltech.edu Fri Jul  5 11:32:37 2002
Date: Wed, 3 Jul 2002 10:49:53 -0700 (PDT)
From: John Zweizig <jzweizig@ligo.caltech.edu>
To: Patrick Sutton <psutton@gravity.psu.edu>
Subject: Re: DMT rev 2.1.0

Hi Patrick,

You're still shooting at a moving target, but apart from that there
shouldn't be much problem. There are a few problems with linux
compilations of LockLoss, ServoMon and glitchMon. If you have any
interest in a linux installation I can give you unofficial fixes for 
these or you can remove them from the Monitors/Makefile. 
Ciao,

John

On Wed, 3 Jul 2002, Patrick Sutton wrote:

> 
> John,
> 
> I'm going to upgrade my machine at Penn State to the "S1" version 
> of the DMT.  Is there anything about 2.1.0 that I should be aware of?
> 
> Cheers,
> 
> -- Patrick
> 
> -------------------------------------------------------------------------
> 
> On Sun, 23 Jun 2002, John Zweizig wrote:
> 
> > Hi All,
> > 
> > I have restarted all the DMT nodes at Hanford, and Szabi will restart the
> > LLO nodes soon.  This brings up version 2.1.0 which will be used in the S1
> > Run. Please check that your monitors are running and producing reasonable
> > data. It is very possible that I have forgotten some monitors or don't
> > have current configurations files at the sites, in which case please tell
> > me. 
> > Ciao,
> > 
> > John
> > 
> > _______________________________________________
> > dmt-discuss mailing list
> > dmt-discuss@tenaya.physics.lsa.umich.edu
> > https://tenaya.physics.lsa.umich.edu/mailman/listinfo/dmt-discuss
> > 
> 
> 

From jzweizig@ligo.caltech.edu Fri Jul  5 11:32:45 2002
Date: Wed, 3 Jul 2002 12:04:31 -0700 (PDT)
From: John Zweizig <jzweizig@ligo.caltech.edu>
To: Patrick Sutton <psutton@gravity.psu.edu>
Subject: Re: DMT rev 2.1.0

The current "latest and greatest" version is tagged gds_2_1_0pre. The 
changes to the monitors are basically:

Add to LockLoss/TSWindow.hh and to ServoMon/TSWindow.hh (in file scope):

#ifdef linux
typedef unsigned long unlong_t;   //---- TYPO: should be ulong_t
#endif

Replace in glitchMon/ChannelInfo.cc 

    #include <ieeefp.h> 

with

    #ifdef linux
    #include <math.h>
    inline bool finite(double x) {return isfinite(x);}
    #else
    #include <ieeefp..h>    //---- TYPO: should be <ieeefp.h>
    #endif

Ciao,

John

On Wed, 3 Jul 2002, Patrick Sutton wrote:

> John,
> 
> Thanks for the warnings.  I would appreciate the unofficial fixes 
> for LockLoss, ServoMon and glitchMon - it's better to have everything.
> Also, I just stuck my head into the CVS repository and found that 2.1
> is not available.  Is there a convenient set of source codes somewhere?
> 
> -- Patrick
    

From jzweizig@ligo.caltech.edu Fri Jul  5 11:32:56 2002
Date: Wed, 3 Jul 2002 16:12:57 -0700 (PDT)
From: John Zweizig <jzweizig@ligo.caltech.edu>
To: Patrick Sutton <psutton@gravity.psu.edu>
Subject: Re: DMT rev 2.1.0

Hi Patrick,

I don't recall having seen that error before, but the External dependency
generation if particularly flaky. Have you tried to run make -C External 
again? sometimes it helps to have everything already compiled.
Ciao,

John

On Wed, 3 Jul 2002, Patrick Sutton wrote:

  
> John,
> 
> I've encountered an unfamiliar error in compiling rev_2.1.0.  At the 
> "make -C External" step I get 
> 
> make[1]: Leaving directory `/opt/dmt/cvs/gds/External/expat'
> make[1]: Entering directory `/opt/dmt/cvs/gds/External/framecpp-0.0.23'
> make[1]: *** No rule to make target `clean'.  Stop.
> make[1]: Leaving directory `/opt/dmt/cvs/gds/External/framecpp-0.0.23'
> make: [include/general] Error 2 (ignored)
> make: *** No rule to make target `../include/sfftw.h', needed by `all'.  
> Stop.
> make: Leaving directory `/opt/dmt/cvs/gds/External'
> 
> I've seen the framecpp complaints before, but not the sfftw.h problem.
> Any ideas?
> 
> Thanks,
> 
> -- Patrick
> 
> ----------------------------------------------------------------------
> 
> Patrick J. Sutton                                   off: 814-863-9599
> Center for Gravitational Wave Physics               fax: 814-863-9608
> The Pennsylvania State University        psutton@gravity.phys.psu.edu
> 
> ----------------------------------------------------------------------