[GE users] StartPVM can't start slave pvmds

Sili Huang shuang at unb.ca
Fri Mar 18 20:05:26 GMT 2005


Hi Reuti,


Reuti> Hi,

Reuti> interesting: in the last couple of days there is again a demand for PVM on this
Reuti> list. I'm just looking into it, to get a tight
Reuti> integration with SGE (the pvmds 
Reuti> have to be children of the shepered, not just to be started by rsh). But
Reuti> anyway:

Reuti> - PVM on it's own is working, when you start the
Reuti> daemons from the pvm console?

Yes, PVM is working in each of the nodes when starting from console.

Reuti> - rsh is possible between the nodes?

Yes, rsh is working across nodes.

Reuti> - What is the content of your /remote/temp_pe_hostfile
$ cat temp_pe_hostfile
v60-n05
v60-n07

The temp_pe_hostfile is just a copy of $TMPDIR/hostfile generated by
$SGE_ROOT/pvm/startpvm.sh, where:

# create pvm_hostfile
# remove column with number of slots per queue
# pvm does not support them in this form
pvm_hostfile="$TMPDIR/hostfile"

# enhance the search path if requested
if [ "x$path_enhancement" != "x" ]; then 
   echo "* ep=$path_enhancement" >> $pvm_hostfile
fi
 
cut -f1 -d" " $pe_hostfile >> $pvm_hostfile



Reuti> Cheers - Reuti

Reuti> Quoting Sili Huang <shuang at unb.ca>:

>> Hi,
>> 
>> I am experiencing problems in getting PVM 3.4 integeraded with Sun
>> Grid Engine 6.0. Does anyone met this problem and solved it before?
>> 
>> StartPVM reports errors when bringing up the slave pvmds:
>> startpvm: Couldn't get all of the 2 requested hosts
>> startpvm.sh: startup failed - invoking cleanup script.
>> 
>> When I am trying to run StartPVM manualy for testing, it reports that:
>> [root at v60-n05
>> root]/remote/general/sge/pvm/bin/lx24-x86/start_pvm -h 2
>> /usr/share/pvm3/lib/pvmd /remote/temp_pe_hostfile
>> /tmp/pvmtmp023986.0
>> start_pvm: enrolled to local pvmd
>> start_pvm: got 1 of 2 hosts
>> start_pvm: 262144 v60-n05 LINUX 1000
>> start_pvm: got 1 of 2 hosts
>> start_pvm: 262144 v60-n05 LINUX 1000
>> start_pvm: got 1 of 2 hosts
>> start_pvm: 262144 v60-n05 LINUX 1000
>> start_pvm: got 1 of 2 hosts
>> start_pvm: 262144 v60-n05 LINUX 1000
>> start_pvm: got 1 of 2 hosts
>> start_pvm: 262144 v60-n05 LINUX 1000
>> start_pvm: got 1 of 2 hosts
>> start_pvm: 262144 v60-n05 LINUX 1000
>> start_pvm: got 1 of 2 hosts
>> start_pvm: 262144 v60-n05 LINUX 1000
>> start_pvm: got 1 of 2 hosts
>> start_pvm: 262144 v60-n05 LINUX 1000
>> start_pvm: got 1 of 2 hosts
>> start_pvm: 262144 v60-n05 LINUX 1000
>> startpvm: Couldn't get all of the 2 requested hosts
>> 
>> Note that there was not orphaned pvmd and /tmp/pvmd.* files in the two
>> nodes.
>> 
>> Thanks.
>> 
>> Best regards,
>> 
>> Wesley Huang
>> 
>> 
>> 
>> 
>> 
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: users-unsubscribe at gridengine.sunsource.net
>> For additional commands, e-mail:
>> users-help at gridengine.sunsource.net
>> 



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



Best regards,
Sili Huang

--
mailto:shuang at unb.ca
University of New Brunswick
Faculty of Computer Science
P.O. Box 4400
Fredericton, N.B. E3B 5A3
Tel(office):  (506) 452-6348


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