Nancy:Talc

From Grid5000

Jump to: navigation, search
See also: TALC | Nancy:Home

Contents

Talc cluster

Talc is a production cluster composed by 134 nodes.
TALC cluster is managed by Grid'5000 Nancy team.

How to get an account

Please find the request form here.

  • The following fields must be filled as indicated (but the other fields must be filled too):
    • manager: lnussbaum
    • site: nancy
    • groups, roles: none
    • privileges: user
  • You are automatically subscribed to the Grid'5000 users' mailing lists: users@lists.grid5000.fr

This list is the user to user or user to admin communication mean to address help/support requests for Grid'5000.
You can make a filter of messages on that list if you want.

How to access and work on Talc

From LORIA network

Image:Terminal.png mypc:
ssh -p 2222 jdoe@access.nancy.grid5000.fr

From external

Use the national access.

Image:Terminal.png mypc:
ssh jdoe@access.grid5000.fr

Then, connect to the frontend for Talc.

Image:Terminal.png access:
ssh talc.nancy

These two steps can be automatized using ~/.ssh/config :

Host access
        HostName access.grid5000.fr
        User jdoe
        Port 22
        IdentityFile /home/doe/.ssh/id_dsa
        ForwardAgent no
Host *.user
        User jdoe
        ProxyCommand ssh access "nc -q 0 `basename %h .user` %p"
        ForwardAgent no
Image:Terminal.png mypc:
ssh talc.nancy.user

Disks utilization

You can use tmp directory during your reservation. Prefer local disk to nfs home during your experiments for better performance. Once your reservation is terminated, tmp will be formatted.

Image:Note.png Note

- access.nancy.grid5000.fr is not the frontend for Talc. It's just the access machine used for the ssh redirection.
- Dont' forget to specify the port 2222 when connecting from LORIA network.

How to access data in Inria/Loria environment

loria1.loria.fr is an access machine hosted on Loria side. That machine can be used to access all services on Inria/Loria environment.

You need to use Ssh ProxyCommand for that purpose.

Adapt following lines for your ~/.ssh/config

  Host accessloria
        HostName loria1.loria.fr
        User jdoe
        Port 22
        IdentityFile /home/jdoe/.ssh/id_dsa
        ForwardAgent no
  Host *.loria
        User jdoe
        ProxyCommand ssh accessloria "nc `basename %h .loria` %p"
        IdentityFile /home/jdoe/.ssh/id_dsa

With that setup, you can now use :

  • Rsync to synchronize your data on Inria/Loria environment and data on your local home on Talc frontend
  • Sshfs to mount directly your data directory on Inria/Loria environment under your local home.

eg:

Image:Terminal.png ftalc:
sshfs -o idmap=user jdoe@tregastel.loria:/users/myteam/jdoe ~/remote_dir

To unmount the remote filesystem:

Image:Terminal.png ftalc:
fusermount -u ~/remote_dir
Image:Note.png Note

The password for your Inria/Loria account will be asked twice if you don't specify a ssh key.

Informations

You can find some useful documentation here :

Network Interlink

Room C105A (Talc)

Production network Talc

Switchs interlink

    gw.nancy    stalc1(T2/1)-stalc1(T1/2)    stalc2(T1/2)    stalc3(T1/1)-stalc3(T1/2)    stalc4(T1/1)-stalc4(T1/2)
      +              +============+               +               +============+               +============+
      |              |            |               |               |            |               |            |
      +--------------+            |               +---------------|------------|---------------+            |
                                  +-------------------------------+            +----------------------------+

Cluster management tools

DrawGantt

Monika

Ganglia

Usage reports (Kaspied)

Special rules for the usage of Talc cluster

  • Only members of the scientific operations of TALC can reserve resources for a given date. Other users have their reservations as soon as resources are available.Prefer submissions (let oar decide when to run the job) to reservations (specifying a time when the job should run).
  • Without permission, any user experiments on the cluster do not must take more than 3840hours for CPU (eg: 10 nodes for 48hours).

In case of problems

The mailing list clustertalc-annonces@loria.fr diffuse all announcements related to the cluster.

To report bugs or problems on the cluster:

How to cite / Comment citer

If you use the Talc cluster for your researches and publish your work, please add this sentence in the acknowledgements section of your paper:

"This work has been partly realized thanks to the support of the Région Lorraine and the CPER MISN TALC project."
"Ce travail a été réalisé en partie grâce au soutien de la Région Lorraine et du projet CPER MISN TALC."a

Personal tools
Wiki special pages