I think I found out what's going wrong but not sure where to fix it.

 

   The sge install files are getting pushed out to the nodes OK

 

   Then the following command is run on the nodes but never completes:

                                

bash -c  cd /opt/sge && ./inst_sge -x > /tmp/oscar_inst_sge.out 2>&1

 

   The contents of /tmp/oscar_inst_sge.out on the node is:

 

TERM environment variable not set.

TERM environment variable not set.

TERM environment variable not set.

 

Checking $SGE_ROOT directory

----------------------------

 

The Grid Engine root directory is not set!

Please enter a correct path for SGE_ROOT.

 

If this directory is not correct (e.g. it may contain an automounter

prefix) enter the correct path to this directory or hit <RETURN>

to use default [/opt/sge] >> 

 

   The last part is coming from the following code in
/opt/sge/util/install_modules/inst_common.sh:

 

      if [ "$SGE_ROOT" = "" ]; then

         while [ "$SGE_ROOT" = "" ]; do

            $CLEAR

            $INFOTEXT -u "\nChecking \$SGE_ROOT directory"

            $ECHO

            eval SGE_ROOT=`pwd | sed 's/\/tmp_mnt//'`

            $INFOTEXT -n "The Grid Engine root directory is not set!\n"
\

                         "Please enter a correct path for SGE_ROOT.\n" 

            $INFOTEXT -n "If this directory is not correct (e.g. it may
contain an automounter\n" \

                         "prefix) enter the correct path to this
directory or hit <RETURN>\n" \

                         "to use default [%s] >> " $SGE_ROOT

         

            eval SGE_ROOT=`Enter $SGE_ROOT`

         done

         export SGE_ROOT

      else

 

  Which appears to be trying to prompt the user for input if $SGE_ROOT
is empty and thus hanging the post_install script.

 

   Where should SGE_ROOT be getting set and is "/opt/sge" the correct
setting for SGE_ROOT on the OSCAR nodes?

 

   Would this question be more appropriate for the oscar-devel list?

 

Thanks,

Brian

 

 

                

 

From: Fessler, Brian [mailto:fess...@lpi.usra.edu] 
Sent: Thursday, April 22, 2010 4:39 PM
To: oscar-users@lists.sourceforge.net
Subject: [Oscar-users] Step 7: SGE post_install hangs

 

Hi, 

 

   I am trying to install Oscar 5.1 RC1 on Red Hat ELS 5.3 on x86_64
hardware. All goes well until running Step 7 where the Complete Cluster
Setup script hangs at:

 

[sge] SGE post_install: Running inst_sge on clients...

 

   The full oscarinstall.log can be seen at http://pastebin.com/WDnecnka
.

 

   There is not a success or failure pop-up window and the original
install window (with the Step buttons) is unresponsive. 

 

   I am not using any third-part packages and things seemed to be going
very well up to this point.

 

   Any ideas as to what's going wrong?

 

   This is very similar to post by James McClearen (2-16-2009) and
Keshetti Mahesh (3-24-2008) but I did not find any resolution to their
issues.

 

Thanks,

Brian

 

------------------------------------------------------------------------------
_______________________________________________
Oscar-users mailing list
Oscar-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/oscar-users

Reply via email to