Information About the InfoLab's Systems

Your home directory

Your home directory is on one of our file servers and is automounted on whichever Linux InfoLab system you log into. If you are not using a Linux system, you will have to mount it or map it as a drive on your system.

Only home directories are backed up! If you have stuff that needs to be preserved on your Windows, Mac or Linux system, you should periodically copy it to your home directory on the file servers.

If you need to refer to your home directory, you should always refer to it as /u/your-login-account-name. Don't do a pwd to find out what it is and then put that into a script or some code. We often need to move home directories around and /u/your-login-account-name will always point to where your home directory is currently located.

Temporary Storage

We have a 8.6TB area for temporary storage, which is automounted on all of the InfoLab's systems at /dfs/tmp/1. This storage is not backed up. If you have stuff that can be easily recreated, please create a directory for yourself at /dfs/tmp/1/your-loging-account-name and put your stuff there, not in your home directory. Please remember to remove your stuff from this temporary storage when it is no longer needed.

Local storage and NFS storage

Your home directory and the Temporary Storage, described above, are NFS mounted on all of the InfoLab systems, so they are available on any system that you access and are mounted as "/dfs/...". There are also some other directories that are NFS mounted across all or some of the InfoLab systems. These directories are also mounted as "/dfs/...".

Each of the systems has some local storage which is mounted as "/lfs/...". The local /tmp directory is always on /lfs/0 on each system. There is some local storage on some of the compute servers which is mounted on /lfs/1. If you need to store stuff in the /lfs/1 partitions, please create a directory for yourself in /lfs/1/tmp. Stuff that is stored in the /tmp (/lfs/0 or /lfs/1) directories is not backed up.

User accounts

We are moving our servers to the CS management which menas that you shall be able to access all our servers with your CS account.

However, things take time and we are still in the transition process which means that some of our servers already use the CS accounts and others are still using the Infolab accounts.

The column login in the following tables will tell you which account to use.

Compute Servers

These servers are used for compute intensive tasks such as graph analyses.

System Name OS Memory [GB] CPU Type Arch CPU Speed [GHz] CPUs Cores Threads Local Storage [TB] Login
Madmax CentOS 6.3 1024 Xeon E7-8837 64 bit 2.66 8 64 64 13 CS
Rulk CentOS 6.3 192 Xeon X5680 64 bit 3.33 2 12 24 26 CS
Hulk CentOS 5.7 192 Xeon X5680 64 bit 3.33 2 12 24 19 CS
Rocky CentOS 5.8 144 Xeon X5667 64 bit 3.06 2 8 16 19 CS
Rambo CentOS 5.6 512 Xeon X7560 64 bit 2.266 4 32 64 11 Infolab
Bruce CentOS 5.6 192 Xeon X5680 64 bit 3.33 2 12 24 20 Infolab
Zarya CentOS 5.3 72 Xeon X5570 64 bit 2.93 2 8 16 20 Infolab
Eel CentOS 5 64 Xeon X5365 64 bit 3.0 2 8 8 3.4 Infolab

You can find more information on the compute servers in the Infolab wiki.

Compute cluster

The compute cluster is running on 36 nodes. Each of the nodes has 32 cores, so we have a total of 1152 cores. We are running the Torque resource manager and Maui scheduler to manage queues and cluster resources.

System Name OS Memory [GB] CPU Type Arch CPU Speed [GHz] CPUs Cores Threads Local Storage [TB] Login Description
ilhead1,ilhead2 CentOS 6.3 64 Opteron 6276 64 bit 2.3 2 32 32 2 CS Job submission nodes
iln01-iln36 CentOS 6.3 64 Opteron 6276 64 bit 2.3 2 32 32 2 N/A Compute nodes
ild1,ild2 CentOS 6.3 64 Opteron 6276 64 bit 2.3 2 32 32 2 CS Development nodes

Here is some information on how to use the compute cluster.

Hadoop cluster

The current Hadoop cluster comprises of 22 nodes with the following hardware configuration.

System Name OS Memory [GB] CPU Type Arch CPU Speed [GHz] CPUs Cores Threads Local Storage [TB] Login
ilc0-ilc21 Ubuntu 10.04 4 Core2 Q6600 64 bit 2.4 1 4 4 3.6 Other

You can find more information on the Hadoop cluster here.

Infrastructure servers

System Name OS Memory [GB] CPU Type Arch CPU Speed [GHz] CPUs Cores Threads Local Storage [TB] Login Description
il-fs-1 CentOS 5.8 12 Xeon E5620 64 bit 2.66 1 4 8 18 CS New file server
Snap CentOS 5.3 8 Core2 Q9400 64 bit 2.66 1 4 4 5.3 Infolab SNAP group's web server
Shark Fedora 8 4 Xeon MP 32 bit 2.0 4 4 8 0.04 Infolab Remote login server
Skate Fedora 8 4 Xeon MP 32 bit 2.0 4 4 8 0.04 Infolab ?
Whale N/A 32 Opteron 865 64 bit 1.8 4 8 8 N/A N/A Crawling server

Printing

The InfoLab's printers are located in 427. Please see the following pages on how to set the printers up in different operating systems:

Creating an InfoLab Web Page for Yourself

You can then refer to the page as http://InfoLab.Stanford.EDU/~<your_account_name>.

Mapping and Unmapping a Drive in Windows

You can map your home directory that is on the linux servers onto your windows machines as follows:

Since we aren't using LDAP, we have to set up yet another password for you. Contact your system administrators and have them create a samba password for you. Then do the following:

     Left Click on "My Computer"
     Right Click on "Map Network Drive"
     In the "Folder" box enter
         \\ILSS\your-unix-account-name
         (Decide for yourself whether or not you want the
                       "Reconnect at login" box checked.)
         Left Click on "Finish"
     In the "Connect to ILSS" window, which should appear enter:
         your linux account name into the "User Name" box
         your samba password into the "Password" box
        (I strongly recommend that you make sure that the
         "Remember my password" checkbox is not checked.)
      
You can unmap a drive as follows:
     Left Click on "My Computer"
     Right Click on "Disconnect Network Drive"
     In the "Disconnect Network Drives" window
         Select the drive(s) that you want to unmap
         Left Click on "OK"
      

Infolab Wiki

You can find more up to date information on the Infolab computer systems in our new wiki: http://snap.stanford.edu/moin/InfolabWelcome


Last updated: Sep 21, 2012