OceanStore/ROC Cluster

The OceanStore cluster was retired August 15th, 2007.

On or after August 15th, 2007, if there is any interest in the continued use of the OceanStore/ROC X cluster, it will be wiped and reinstalled to bring it into compliance with the UC Berkeley Minimum Security Standards for Networked Devices. If there's no interest, the cluster will be shut down to conserve electricity.

Since there was never an auto-deletion policy on the X cluster, the /scratch and /work filesystems are littered with old data. Please verify you have retrieved any important data before this date.

Furthermore, the fileserver (xm3) which provides /work to the X cluster is on the verge of death and will be shut down. If you have essential data, we'll work with you to back it up or find it a new home.

Those of you whose EECS accounts have recently expired have been granted extensions to your Millennium accounts until Aug 15th to facilitate this cleanup.

X Cluster

The X Cluster currently has 40 compute nodes; x{1-40}.Millennium.Berkeley.EDU

  • 42 IBM xSeries 330 (X Cluster)
    • 2x 1.0 GHz Intel Pentium_III CPUs
    • 1.5GB ECC PC133 SDRAM
    • 2x 36GB IBM UltraStar_36LZX Hard Disk
    • Intel PRO/1000F Server Adapter
    • Myricom Myrinet PCI64A Adapter (LANai 7.2)
  • 1TB Fiber Channel Storage
  • 2 Foundry_Networks BigIron_4000
  • 1 Foundry Networks ServerIron

Tue Aug 29 2006: x21 is offline due to hardware failures. Sep 2006: x28-x42 are reserved by the fall 2006 RAD Lab class.

IBM Cluster

The IBM Cluster currently has 8 compute nodes; ibm{1-8}.CS.Berkeley.EDU

  • 8 IBM Netfinity 4000R 1U rackmount PC (IBM Cluster)
    • 750 MHz Intel Pentium_III CPU
    • 1GB ECC SDRAM
    • 1 or 2 18GB IBM SCSI Hard Disk
    • Intel pro100 ethernet

Jan 1 2007: The IBM cluster has been decomissioned. Fri Dec 22 2006: Fhe entirety of the IBM cluster is offline for the holiday break, if not forever. Fri Oct 13 2006: ibm6 is offline due to hardware failures.

Using the Cluster

Access is currently limited to active members of the OceanStore and ROC projects. Exceptions will be made with consent of the projects' PI's.

This cluster is a shared resource. You can view the current use of the machines by running gstat or viewing the Ganglia graphs.

gstat provides an ordered list of available machines. Machines with a load >= 2 are fully loaded and should be avoided until current jobs have completed.

Other Cluster Resources

  • FAQ
  • Reservation_System
  • Documentation
  • Mailing_List & Archives

Filesystems

While you should be able execute jobs from your EECS department home directory, we strongly suggest that you launch all jobs from a /work/$user or /scratch/$user directory. Before executing a program, copy all binaries and data files into your /work or /scratch directory, cd into that directory, and execute from there. This avoids putting unnecessary load on EECS department fileservers, which are sometimes unable to handle many simultaneous mount requests.

Note: /work has NO deletion policy. /work is not for long-term storage and is never backed up. /scratch has NO deletion policy; please clean up after yourself. /scratch is RAID 0 local to each machine. /scratch is never backed up.

 
oceanstore.txt · Last modified: 2008/09/30 20:04 by mhoward
 
Except where otherwise noted, content on this wiki is licensed under the following license:CC Attribution-Noncommercial-Share Alike 3.0 Unported
Recent changes RSS feed Donate Powered by PHP Valid XHTML 1.0 Valid CSS Driven by DokuWiki