[torqueusers] problem with: set queue cfq route_destinations
guillaume.alleon at laposte.net
Wed Aug 10 02:42:54 MDT 2005
Yes, I do confirm that (at least with 1.2.0p5) you don't need to place
the routing destination to the nodes file.
I have a machine running a pbs_server without nodes file routing jobs to
a remote pbs_server and this is working
like a charm.
The only point is to modify the iptables on this remote server so that
mom nodes can "scp" back to the submitting
master (so called point 7 in Stewart's answer). Could someone comment on
this point this I am not familiar with
Gerson Galang a écrit :
> You might not be seeing that problem anymore because as what Etienne
> Gondet said, torque-1.2.0p4 has already integrated the patch below
> which fixes this problem even if you haven't added the routing
> destination to the nodes file.
> Garrick Staples wrote:
>> On Mon, Aug 08, 2005 at 03:02:28PM -0400,
>> Stewart.Samuels at sanofi-aventis.com alleged:
>>> When digging through the code, what I found was that pbs_server
>>> (when dealing
>>> with the routing of jobs) actually uses the "pbsnodes" command to
>>> get a list
>>> of valid destinations to which it can route. In other words,
>>> validates the "host" portion (specified in the form "@hostname")of
>>> the names
>>> listed in the "route_destinations" attribute with the names of the
>>> listed in its own $PBS_HOME/server_priv/nodes file.
>>> So, in your case, simply place "other.host" in your nodes file on
>>> the system
>>> in which you are trying to transfer from. The pbsnodes command must
>>> show the
>>> remote node listed in your $PBS_SERVER/server_priv/nodes file as
>>> "free". For
>>> this to occur, the $PBS_SERVER/mom_priv/config file on the remote
>>> system must
>>> contain a $clienthost entry for the sending node. If the remote
>>> server is a
>>> master node for a cluster, run a mom on the master node that
>>> contains the
>>> sending server in its $PBS_HOME/mom_priv/config file. The compute
>>> nodes for
>>> the cluster do not require this entry.
>> I've been trying this and can't seem to duplicate it. I have no problem
>> submitting and routing to remote servers without entries in
>> The only requirement that I can find is that the submitting user and
>> host must pass
>> userok() (add the submitting host to the receiver's /etc/hosts.equiv).
>> This is with torque-1.2.0p5-snap.1122411769 which is almost the same
>> thing as
>> the 1.2.0p5 release.
>> torqueusers mailing list
>> torqueusers at supercluster.org
> torqueusers mailing list
> torqueusers at supercluster.org
More information about the torqueusers