[GE users] PVM tight integration

Reuti reuti at staff.uni-marburg.de
Tue Apr 11 16:56:35 BST 2006


Am 11.04.2006 um 17:26 schrieb davide cittaro:

> Hi Reuti, I solved the first issue: I can start pvmd on all slave
> nodes, I modified the startpvm.sh in this way:
>
> 133a134,136
>> # add pvmd path to pvm_hostfile
>> echo "* dx=/usr/share/pvm3/lib/pvmd" >> $pvm_hostfile
>>
>
> so that pvm knows where to find the pvmd executable. Apparently it
> doesn't work with loose integration (there I have to specify the
> PVM_DPATH as -v variable).
> Although I can run pvm, it still doesn't create pvmd.$ID and sockets
> in the right place... looking at the errors, it searches them in
> /tmp/$JOB_ID.$QUEUE but they are created in /tmp. I assure that I
> don't modify TMPDIR values... Do you use a .pvmrc file to specify some
> variable?

No.

With all this things, maybe it would be worth to try a self compiled  
PVM in your ~, as I implemented all the things this way. Maybe there  
is something compiled into your distribution.

-- Reuti


> d
>
> ---------------------------------------------------------------------
> 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