[GE users] Start execd script don't work - /tmp/exec* says, can't create directory "bnode": No such file or directory

Daniel Templeton Dan.Templeton at Sun.COM
Sun Dec 16 17:47:37 GMT 2007


    [ The following text is in the "ISO-8859-1" character set. ]
    [ Your display is set for the "ISO-8859-10" character set.  ]
    [ Some special characters may be displayed incorrectly. ]

At some point in the past, I had a problem installing the execution 
daemon because of permission problems on the spool directory.  To get 
around the problem, I had to go to the spool directory and create a 
directory named for the execution daemon (bnode, in this case) and set 
the SGE user (sgeadmin in most cases) as its owner.

Daniel

Rayson Ho wrote:
> execd needs to create a sub-directory for it to store the spooling information.
>
> Looks like execd couldn't create the sub-directory, may be a
> permission problem, or some NFS problem??
>
> Note that you can also create a local spool directory:
> http://gridengine.sunsource.net/howto/nfsreduce.html
>
> Rayson
>
>
>
> On Dec 15, 2007 9:58 PM, amanyus <amanyus at gmail.com> wrote:
>   
>> Hello all users! Need some expertise here. The init.d script does not work
>> for my x86 Linux.
>>
>> #  cat /tmp/exec*
>> 12/16/2007 05:52:35|execd|bnode|C|can't create directory "bnode": No such
>> file or directory
>> 12/16/2007 05:53:24|execd|bnode|C|can't create directory "bnode": No such
>> file or directory
>>
>>
>> Investigation:
>> [root at bnode lx24-x86]# ps -ef | grep sge
>> root      3292  2708  0 10:53 pts/1    00:00:00 grep sge
>> [root at bnode lx24-x86]# pwd
>> /opt/n1ge6/bin/lx24-x86
>> [root at bnode lx24-x86]# ./sge_execd
>> critical error: can't create directory "bnode": No such file or directory
>> <<< HELP
>> [root at bnode lx24-x86]#
>>
>>
>> Information:
>>
>>  bnode:/opt/n1ge6# ./util/arch
>>  lx24-x86
>>
>> From /etc/init.d/sgeexecd:
>>
>> for i in $*; do
>>  if [ "$i" = start ]; then
>>  startup=true
>>
>> ...
>>
>>
>>
>> if [ "$startup" = true ]; then
>>
>>    # execution daemon is started on this host!
>>
>>       echo "   starting sge_execd"
>>       exec 1>/dev/null 2>&1
>>       $bin_dir/sge_execd
>>       [ $? -eq 0 -a -d /var/lock/subsys ] && touch /var/lock/subsys/sgeexecd
>>       exit 0
>> else
>>    if [ $execd = true ]; then
>>       # Shutdown execution daemon
>>       GetExecdSpoolDir
>>       execd_spool_dir=$EXECD_SPOOL_DIR/$UQHOST
>>
>>       # Send SIGTERM to execd
>>       echo "   Shutting down Grid Engine execution daemon"
>>       Shutdown sge_execd $execd_spool_dir/execd.pid
>>       ret=$?
>>       if [ -f /var/lock/subsys/sgeexecd ]; then
>>           uid=`$utilbin_dir/uidgid -uid`
>>           if [ "$uid" = "0" -a "$ret" = "0" ]; then
>>              rm -f /var/lock/subsys/sgeexecd >/dev/null 2>&1
>>           else
>>              echo "Can't shut down execd!"
>>              exit 1
>>           fi
>>       fi
>>       if [ $softstop = false ]; then
>>          # Send SIGTERM to all shepherds
>>          for jobid in `ls $execd_spool_dir/active_jobs`; do
>>             echo "   Shutting down Grid Engine shepherd of job $jobid"
>>             Shutdown sge_shepherd $execd_spool_dir/active_jobs/$jobid/pid
>>          done
>>       fi
>>    fi
>> fi
>>
>>
>>
>>
>>     
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe at gridengine.sunsource.net
> For additional commands, e-mail: users-help at gridengine.sunsource.net
>
>   

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe at gridengine.sunsource.net
For additional commands, e-mail: users-help at gridengine.sunsource.net




More information about the gridengine-users mailing list