[GE users] pe_hostfile, how is created?

Reuti reuti at staff.uni-marburg.de
Tue Apr 4 20:00:09 BST 2006


Am 04.04.2006 um 17:56 schrieb davide cittaro:

> Hi all, I would like to know how the $pe_hostfile variable is  
> treated/generated
> According to the man pages is the path to the file with the machine
> names, but how the file is it created? And who creates it?
> I ask this because I have machines with long hostnames, such as
>
> ia32.bioinfo.ifom-ieo-campus.it
>
> where ifom-ieo-campus.it is the domainname and
>
> $ /usr/bin/hostname
> ia32.bioinfo
>

This looks similar to the Rocks installation, where you always get  
this .local - but I wonder, whether it's really allowed by  
definition, that a hostname may contain dots, or whether your  
"bioinfo" isn't already treated as part of the domainname. In the  
end, it wouldn't be easy in your case to decide from the FQDN where  
the hostname ends.

You can of course specify during installation, that SGE should use  
the FQDN - then you should always get the long names.

> but when I try loose lam integration with rsh (first step and attempt)
> I have SGE that tries to lamboot machines only with the first part of
> the hostname (i.e. ia32). I can set these *short* hostnames in
> /etc/hosts, but I also would like why SGE doesn't like a punctated
> hostname (ia32.bioinfo). Can anybody give me a hint?
> Also, a general question, the startSOMETHING.sh/stopSOMETHING.sh
> script should be installed on all machines? I have noticed that

Yes, or put it somewhere in your $SGE_ROOT. It must be installed on  
all machines, which might become a headnode of your parallel job.  
Only there the {start,stop}SOMETHING.sh are executed - but this can  
be of course any machine in your cluster, so it must be available on  
all machines.

-- Reuti


> modifying one startlam.sh script, so that the machines files contains
> the right hostname, makes everything work even if the other machines
> have a *wrong* script.
>
> Thanks
>
> d
>
> --
> dawe
> http://dawe.ilbello.com
> ---
> "Prediction is very difficult, especially if it's about the future." -
> Niels Bohr
>
> ---------------------------------------------------------------------
> 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