Cluster Resources News
Moab Access Portal® Source Code Available
To improve Moab's ease-of-use and expand administrator control, Cluster Resources released Moab Access Portal source code to its customers and partners. This new benefit lets organizations configure the Access Portal to accommodate unique site-specific needs. For example, organizations using Moab can now create automated batch submission on the Access Portal so end users don't have to learn batch codes, which decreases training requirements for administrators.
With available source code, organizations can also integrate Moab Access Portal with their existing Web portals. End users can use the Web portal they are already familiar with and organizations can consolidate programs to simplify management.
Moab Access Portal source code is a value add of Moab Cluster Suite®, Moab Grid Suite® or a Cluster Resources' support contract. Please contact Cluster Resources at (801) 873-3400 to learn more about downloading Access Portal source code.
Cluster Resources releases TORQUE 1.2.0p6
Cluster Resources, in collaboration with the TORQUE* user community, recently released version 1.2.0p6 of TORQUE Resource Manager – a resource manager built on the original OpenPBS, which includes over 1,200 patches to make it more scalable and stable. The newest TORQUE version enables expanded configuration and management.
TORQUE 1.2.0p6 enables MOM configuration with features contributed by USC and OSC. New features include:
-Improved rapid start-up
-Direct support for external scheduler extensions
-New 'shell-us-arg' patch (OSC)
-File recovery failure messages (USC)
TORQUE 1.2.0p6 also includes a reworked MOM-server state management system. The improved management system lets TORQUE perform tasks such as:
-Consolidate node state changes
-Force a health check script when an update state is forced
-Improve pbsnodes 'offline' management
Cluster Resources wishes to thank all those who continue their efforts in contributing to TORQUE Resource Manager. Thanks to these contributions, TORQUE has become increasingly stable, providing a reliable open source resource manager. Those who wish to contribute may join the TORQUE user list at http://www.clusterresources.com/mailing.shtml
TORQUE 1.2.0p6 is now available for download at http://www.clusterresources.com/products/torque/
. For professional support information, visit http://www.clusterresources.com/techsupport.shtml
*TORQUE Resource Manager includes software developed by NASA Ames Research Center, Lawrence Livermore National Laboratory, and Veridian Information Solutions, Inc. Visit www.OpenPBS.org for OpenPBS software support, products, and information. TORQUE is neither endorsed by nor affiliated with Altair Grid Solutions, Inc.
New Documentation Features
Cluster Resources released new PDF and web based documentation for several products in the past month. The PDF documentation allows users to locally browse and print manuals for convenient offline review. The web based documentation includes numerous content, formatting, and organizational enhancements recommended by the community. Cluster Resources is also releasing WIKI-based online documentation in the coming weeks to allow greater community involvement in recommending best practices, assisting each other with usage issues, etc. The new documentation is available to the community at the following locations:
Moab 4.2.2: http://clusterresources.com/moabdocs/MoabWorkloadManagerAdminManual.pdf Maui 3.2.6: http://clusterresources.com/moabdocs/MoabWorkloadManagerAdminManual.pdf Moab 4.2.3: Available to customers and partners with their login account information. TORQUE 1.2.0: http://www.clusterresources.com/torquedocs20/
Cluster Resources is continually trying to improve documentation and is happy to receive suggestions from the user community. Please e-mail suggestions to email@example.com.
New Maui Snapshot
Cluster Resources recently released a new Maui snapshot that fixes a number of internal job management issues, improves fault handling and logging, and updates a number of communication protocols to allow better interaction with peer services.
This snapshot also follows a new version formatting that includes explicit snapshot version information at runtime. The formatting means that administrators can issue a "maui --version" or "showq --version" and see both the Maui release and snapshot version info for the installed package. Based upon feedback from the user community, the formatting change is intended to let administrators know the exact version of Maui they are using. Both TORQUE and Moab previously used this version formatting system.
Please continue to submit patches, suggestions, and comments firstname.lastname@example.org
so we can continue to enhance this product.
The latest Maui 3.2.6 patch 14 snapshot is now available for download at Maui's Download Page
. Cluster Resources also offers technical support for Maui. Learn more at www.clusterresources.com/techsupport
Cluster Resources Training
Cluster Resources' technical team wishes to thank the participants of August's training. Turnout was exceptional, with attendance exceeding 150 percent of anticipated levels. Because there was not room for everyone who wanted to participate, Even with overbooking, Cluster Resources was not able to address all demand and will host another free training session in January.
If you are interested in participating in the January training, please contact Amber Webb at email@example.com
at your earliest convenience to reserve your slot.
Cluster Resources and Scyld Software Form Alliance
Cluster Resources and Scyld Software recently announced an agreement that extends Scyld's Beowulf Linux Clustering solution to include Scyld Taskmaster Suite, a workload and policy management toolset based upon Moab technologies.
The suite, which is built around Moab's architecture, will be officially available in the next few months. The suite provides end users and system administrators with a sophisticated workload policy manager, advanced scheduling and a highly intuitive graphical user interface, accessible locally for administrators or through a remote Web portal for end users.
Cluster Resources and Scyld Software created Scyld TaskMaster Suite to offer customers combined benefits from both companies. Cluster Resources’ customers have experienced significant performance improvements with system utilization levels in the 90 percent to 99 percent range, and Scyld Software’s customers have experienced total cost of ownership savings approaching 50 percent compared to alternative clustering approaches.
"Scyld Beowulf is known for dramatically simplifying the deployment and management of Linux clusters," said Marco Annaratone, general manager of Scyld Software. "The new policy-based management capabilities of Scyld's TaskMaster Suite, with its graphical user interface and Web-based access, allows engineers, scientists and researchers to spend more time focused on critical development efforts.”
To learn more about Scyld TaskMaster Suite, visit: Scyld Taskmaster Suite's homepage
and to learn more about its underlying technology, Moab Cluster Suite, go to: Moab's home page
Cluster Resources Hosts Exhibit at SCl05
Visitors to the annual Supercomputing 2005 (SCl05) conference can visit with Cluster Resources'TM Developers and staff and test Cluster Resources' software live. Cluster Resources will host an exhibit at the conference held in Seattle, Wash. Nov. 15-17 2005.
Guests can ask questions and provide feedback on MoabTM, Maui or TORQUE to a team of technical and product managers. The conference gives visitors a chance to discuss unique and detailed cluster or grid management challenges and brainstorm possible solutions with developers. There is expected to be free evaluation copies of Cluster Resources' software that anyone can take and use on their clusters or grids.
Cluster Resources' booth, number 2036, is located at the south end of the Washington State Convention and Trade Center. To learn more about the SCl05 conference, visit the Web site at http://sc05.supercomputing.org/home.html
. To find a summary of the software solutions Cluster Resources will present, visit http://www.clusterresources.com/downloads/
Cluster and Workload Management Training
Cluster Resources is offering free training on cluster and workload management during the last week of August. One day classes will be held Monday Aug. 22 through Friday Aug. 26 via VNC/Teleconference sessions and will cover topics such as sources of cluster productivity loss, holistic cluster management and end user and administrator empowerment.
Each days' training will be broken into two three-hour sessions and will include live configuration, live troubleshooting, live demos, and interactive usage of graphical and web-based user and administration tools from Cluster Resources' technical team.
To request an invitation to the conference, please email Cluster Resources' Outreach Director at firstname.lastname@example.org. Each class is limited to 20 people and each organization may allow as many people as desired to attend. First preference will be given to paying customers, but others will be invited as space allows.
The outline of training is as follows:
Morning Session 9:00 AM MST to 12:00 AM MST (Mountain Standard Time)
- Overview of the Modern Cluster
- Cluster Evolution and the Source of Cluster Productivity Losses
- Detailed Analysis of Political, Resource, Middleware, and
- Moab Workload Manager Architecture
- Addressing Fairness Issues
(strategies for priority, usage limit, fairshare, and allocation
- Enabling QoS and Service Level Agreements
(using job deadlines, QoS, rollback reservations, and other features)
- Distributing Resources
(using reservations, affinity, node sets, queues, and partitions for
optimal efficiency and responsiveness)
- Reducing the Downtime Associated with Downtime
(automated failure recovery, system reservations, interleaved
maintenance, and high availability)
- Improving Resource Monitoring and Reporting
(native resource managers and generic features, consumable resources,
metrics, and events)
- Holistic Cluster Management
(integrated management of network, storage, and license resources)
- Peer to Peer Grids
(enabling highly flexible grids in two minutes or less - Live Demo)
Afternoon Session 1:00 PM MST to 4:00 PM MST (Mountain Standard Time)
- Automating Responses to Environments Conditions
(dynamically modified policies, triggers, reservations, notifications
and other actions in response to failures, altered states, and
- Trouble Shooting and Diagnosing Performance
(using integrated Moab analysis features to evaluate everything from
fairness policies to data staging)
- Capacity Planning
(using advanced statistics, profiling, and simulation capabilities to
identify cluster bottlenecks and predict future resource needs)
(using job profiles to enable Moab to 'learn' best application
placement, tuning reservation policies, and identifying policy based
- End User Empowerment
(Moab Access Portal overview, personal reservations, and improved
workload placement and start estimation tools)
- Tools for Training and Empowering Administrators
(Moab Cluster Manager overview including integrated documentation and
- Tools for Managers
(Automating professional report and chart generation displaying
performance and usage for users, accounts, nodes, and clusters)
- New Features of Moab 4.2.4
(hosting center support - Live Demo, virtualized resources, dynamic
jobs, optimized data staging, application farms, identity management
interface, and other new features)
- A Look to the Future
(a brief description of the future roadmap for Moab Cluster Suite and
Moab Grid Suite)
Announcing Cluster Builder
Cluster Resources, Inc. and LinuxHPC.org announced the release of Cluster Builder.org
— a new Web site highlighting a broad spectrum of technologies and high performance computing (HPC) information.
Cluster Builder acts as an organized repository of first- and third-party content relating to supercomputing technologies that helps alleviate the research burden of sifting through millions of search engine results. The Web site is organized into software categories such as operating systems, resource managers, compilers, etc., as well as hardware categories such as processors, switches and interconnect standards. Cluster Builder includes both open source and commercial products.
To further aid in the research, Cluster Builder directs evaluators to LinuxHPC's free Request-For-Quote service. Purchasers can use this service to complete one quote form, which is sent to multiple vendors, instead of completing a form for each individual vendor.
Data Staging for Moab Coming Soon
The next release of Moab Workload Manager (v4.2.4), scheduled for release in August, will include intelligent handling of job data staging. This feature can increase compute, networking and overall cluster performance as much as 40 percent, depending on the workload and cluster infrastructure.
Several levels of data staging functionality will be supported depending on the amount of control and information available to Moab. The most basic level of data staging will schedule a job only after any input files are fully staged. The most advanced level of data staging, which will be available in the subsequent release, will fully schedule the data staging operation so that it completes just before the scheduled start time of the job. Cluster Resources is currently Beta testing the new functionality with several partners.
Moab Access Portal For Grid Now In Alpha Stage
Moab Access Portal for Grids provides expanded capabilities for Moab Grid Suite. Now in an Alpha stage, this new version of Moab Access Portal allows end users to not only manage their workload on the cluster level, but also to run workload on grid resources. Running grid jobs is as easy as selecting whether a job should be placed in a grid queue or run on a particular cluster. Grid resources appear to end-users as virtual nodes and can be monitored just like local resources.
Moab Access Portal for Grids provides a seamless transition for cluster-centric users to move to the more powerful grid environments. To view the cluster version of Moab Access Portal, go to http://www.clusterresources.com/eval
and click on the Portal Online Demo link at the bottom of the page. Expect to see Moab Access Portal for Grids released to the public sometime during the third quarter of this year.
Job Templates Now Supported in Moab Cluster Manager
The latest Moab Cluster Manager version 2.2.1 now has support for job templates. This convenient new feature allows users to save their job submission settings for later use. Using job templates, a user can focus on the job itself and ignore the repetitious and often tedious task of entering the same job submission script multiple times.
Users can save personal job templates that are available exclusively to the creator. Moab Cluster Manager also supports global job templates. When administrators or users create global job templates, the job submission settings are saved to a global location where other users can access the templates. This allows them to carefully set up a job submission template that dozens of other users can simply load and apply. Users can modify any settings and then submit the job.
Maui Release 3.2.6p13 Now Available
The newest release of Maui, version 3.2.6 patch 13, has been released to the public.
This release includes numerous usability enhancements and bug fixes.
The biggest single enhancement incorporates the latest Department of Energy S3 (scalable systems software) infrastructure.
This infrastructure allows encryption, compression, improved XML management, improved internal diagnostics and other features.
Moab Cluster Manager Evaluation Software in Beta Phase
Cluster Resources will soon be releasing Moab Cluster Manager v 2.2.2.
For those interested, a beta evaluation (v 2.2.1) of Moab Cluster Manager is available from Cluster Resources' Web site.
The new features in the evaluation allow users to visually modify reservations, graph current and historical node usage (according to node attributes such as architecture, OS and memory) and visually diagnose credential access rights along with cluster and credential statistics.
To download the evaluation, visit our software evaluation page and follow the instructions.
The evaluation connects to Cluster Resources' online demonstration cluster and allows users to explore the functionality as they desire.
Moab Access Portal v 1.2.2 Released
The release of Moab Access Portal v 1.2.2 includes several new features and enhancements.
In addition to the original job submission capabilities, the portal now has a basic job submission screen that enables users to create jobs quickly, without needing to know all the details of a more advanced job.
Also, there are two features available in the portal release to assist users in launching jobs.
First, users can now use pre-existing command files to launch jobs on the cluster.
Second, by viewing the newly integrated estimated start time table, users can make better decisions on processor count and duration options to submit in order to achieve the fasted job start time.
In the portal, job submission settings can be saved in templates (also known as profiles) on both a global and a user level.
These can be loaded at any time, easing the submission process.
Other enhancements include support for other languages (internationalization), an easier installation and a simplified configuration.
ClusterWorld Summit 2005
Cluster Resources will present next-generation utility-based computing and discuss Moab's ability to enable hosting services at the ClusterWorld Summit 2005.
Cluster Resources will primarily focus on how this technology is implemented in two existing client solutions:
A demonstration of the ability of Alexa Internet, a subsidiary of Amazon.com, to take existing high performance computing services and extend access to them across a broader audience.
Alexa Internet's data mining services provide intelligent searching of multiple copies of the Internet that Alexa has constantly updated and stored locally.
A joint presentation with IBM demonstrating that any site can receive instant access to hosted resources, potentially eliminating all HPC backlogged work or significantly reducing queue wait times.
The presentation will take place at ClusterWorld's third annual summit.
The conference provides information on the latest cluster technologies helping organizations stay competitive in High Performance Computing.
Announcing TORQUE 1.2.0p2 and Maui 3.2.6p12 Releases
Cluster Resources is pleased to announce that TORQUE 1.2.0p2 is now officially released and may be downloaded from the TORQUE download page
This release of TORQUE includes:
- Enhanced network failure messages
- A fixed tracejob tool to only match correct jobs (Courtesy of WESTGRID)
- A modified reporting of Linux availmem and totmem to allow larger file sizes
- Fixed pbs_demux for OSF/TRU64 systems to stop orphaned demux processes
- Added dynamic pbs_server loglevel specification
- Added intelligent mom job stat sync'ing for improved scalability (Thanks to USC/CRI)
- Added mom state sync patch for dup join (Courtesy of USC)
- Added spool dir space check (Thanks to MAINE)
The 2005 Maui release marks the 10th anniversary since its creation.
It can be downloaded from the Maui download page.
This release of Maui includes:
- Improved security
- Improved fault tolerance
- Improved resource limit enforcement policies
- Improved SLURM and LSF support
- Improved workload prioritization and diagnostics
Both Cluster Resources' community and professional software projects are growing at a torrid pace with significant growth among Fortune 1000, general commercial, government and academic organizations.
Cluster Resources wishes to recognize and thank all of the above listed organizations for their contributions to this release, as well as all those actively participating in our TORQUE and Maui user communities.
New Partnership with LinuxHPC.org
Cluster Resources recently announced a new partnership with LinuxHPC.org, an online site that provides news, technical information and a quoting service for purchasing Linux clusters.
LinuxHPC.org assists organizations by providing one central location from which they can request quotes for cluster hardware and software from multiple vendors.
LinuxHPC.org gathers information regarding the number of processors, amount of memory, needed accessories etc.
They then send the submitted information to the requested vendors.
This quoting service is designed to save organizations time, simplify the quoting process and ensure they receive back the desired number of quotes.
Cluster Resources teamed with LinuxHPC.org to assist in verifying and facilitating the request for quotes.
Further, Cluster Resources acts as an advocate to the submitting organization by following up with vendors to ensure their responsiveness.
Cluster Resources plans to provide additional value to the site by creating a cluster building worksheet, which will make decisions for hardware and software, identify key factors in architecture, find appropriate software elements and provide links to corresponding Web sites.
For those interested in building or purchasing a cluster, please visit LinuxHPC.org's Cluster Quote Request Form.
Those interested in contributing suggestions, solutions or product information to the cluster building worksheet may contact Cluster Resources at: email@example.com.
Cluster Resources featured in ClusterWorld Magazine, March 2005
The Moab Cluster Suite
"Rather than adding yet another option to the broad array of commercial and open source resource managers out there, Cluster Resources has released a product to aid in managing the managers.
If you have a single cluster, even if you already use a resource manager, Moab can add a GUI, remote management, and additional scheduling capabilities."
"Ultimately, there are four things I really liked about this software that make me want to keep using it in place of command line management of my resource managers:
- The interface makes sense and gives you new capabilities
- It's possible to tweak your scheduling without interrupting the running resource manager
- You have the ability to project the impact of changes before implementing them
- The reporting is about the best I've seen; the graphical reports are easy to understand, and the information available, particularly for completed jobs, is comprehensive."
Moab Workload Manager
"The Moab Workload Manager is the child of the Maui scheduler, and Maui's already fairly sophisticated scheduling algorithms have been taken much further in Moab."
Moab Cluster Manager
"A good GUI provides an intuitive way to navigate a complex series of options without in-depth knowledge of syntax, and ideally provides you information to make decisions you wouldn't have without the GUI.
The Moab Cluster Manager achieves this goal."
"The graphics are clean, the information is logically presented, and I was particularly impressed by some of the new capabilities."
"If you feel that your resource manager is a mysterious black box you simply hope you don't break, you should take a long hard look at Moab."
TORQUE Version 1.2 is officially released
The latest release of TORQUE
provides multiple reliability and scalability enhancements.
Among these are better MOM handling of network and system failures, and improved data management under adverse conditions.
For a complete list of fixes and features see the CHANGELOG in the latest release.
Working with partner sites, Cluster Resources, Inc. has also enhanced the TORQUE Resource Manager to provide several workarounds for OS-level client-server communication failures and memory reporting issues in Mac OS X.
The latest release has solved all known issues associated with these failures.
Great progress has been made on all fronts and a huge thanks goes out to the many sites contributing patches and suggestions.
TORQUE now contains more than 600 contributed patches, enhancements, and design improvements, and the contributions are continuing unabated.
Testing of this latest release has gone extremely well, and Cluster Resources continues to get a very positive response from all sites and platforms.
Further thanks go out to the many TORQUE users who are making excellent suggestions for the road map and helping each other with usage and other issues.
TORQUE 1.2.0 has been a major stability release.
If any issues are uncovered, please let us know via the mailing list or Bugzilla.
The next release will be a development release containing multiple enhancements -- support for new platforms and capabilities (i.e. Altix an CPUSets), improved diagnostics, and increased usability.
CTO David Jackson speaks at GlobusWORLD 2005
Dave Jackson presents the resource management capabilities of Moab Grid Scheduler at GlobusWORLD 2005
, and discusses:
- Moab Grid Scheduler's ability to pull together resources of varying types and apply them in a focused effort,
- The underlying co-allocation problem and the methodologies employed within Moab Grid Scheduler to allow co-allocation of resources along both time and space dimensions, and
- The architecture employed by Moab Workload Manager to make arbitrary resources available for reservation and scheduling within a grid environment.
The GlobusWORLD conference, begun in 2003, is considered the most important event of the year addressing new grid technologies and applications.
The conference features training, presentations, and panels given by leaders in the industry.
||Maui Consortium BOF to be held at SC2004 at 12:15 on Nov 9, in Room 324
||Cluster Resources releases Moab Workload Manager 4.2.0
||Cluster Resources creates evaluation versions of Moab Workload Manager
||Moab Access Portal in Beta
Cluster Resources Joins Novell in BrainShare Keynote on High Performance Computing.
||Cluster Resources Software usage grows nearly 300% over previous year.
||TeraGrid selects ClusterResources for Cluster and Grid Management services.
||Moab Grid Manager selected for 1000 proc NIH-CHPC Opteron Cluster.
||Torque patch 4 released. Over 800 downloads of Torque in last two months.
After more than a year of development, alpha testing begins for Moab Cluster Manager.
MCM is an advanced graphical Cluster Management Tool providing resource management and scheduling configuration, tuning, analysis, and trouble-shooting capabilities.
||Maui 3.2.6 available for early access users
||SSS Job Specification 2.0 released
||Maui 3.2.6 begins beta test
||Initial SGE support enabled
||Maui 3.2.5 released for production use
||MauiCon hosts SC 2002 Maui BOF
||Maui 3.2.5 Scalability demonstrated to 160,000 processors (simulation demonstrates effective scheduling for hundreds of jobs on massive cluster)
||Maui usage continues to grow, this year over 5,000 downloads, 1 million page hits!
||Maui-LSF interface developed (see LSF Integration Guide)
||MauiCon supports API extensions for LL 3.1/PBSPro 5.2
||Maui Admin manual now available in PDF (see documentation)
||Clubmask BProc based cluster building distribution released with integrated support for Maui (see ClubMASK Homepage)
||Efforts begin in conjunction with the Ohio Supercomputing Center to integrate Silver and Globus providing intelligent scheduling of the 'Cluster Ohio' grid.
||Maui usage grows. Over 2000 downloads of Maui in the last year.
||The U.S Department of Energy announces selection of Maui and Silver as the scheduling and metascheduling components of their 'Scalable System Software Initiative' to develop advanced systems software to support the next generation of massive clusters. This 5 year effort will incorporate significant enhancements in the areas of scalability, security, and scheduling functionality.
||Two new resource manager projects started using Maui as a primary scheduler. Members of the University of Pennsylvania's Liniac project begin work on a BProc based system while the Parallel Research Group at Thailand's Kasetsart University works on their Scalable Cluster Environment project.
||Work on the Maui Scheduling chapter of the new 'How to Build a Beowulf' book is completed. Look for it in stores in a few months.
||CHPC begins work on Compaq/RMS-Maui port
||Maui research presented at International Parallel and Distributed Processing Symposium (IPDPS) 2001. Additional papers on Maui research to be presented soon at the IEEE Sigmetrics Advanced Scheduling Workshop, the International Conference on Parallel Processing, and the NCSA 'Linux Revolution' conference.
||NCSA runs Maui on 512 node PBS Linux cluster.
||Maui packaged in RPM format for use in NPACI Rocks HPC Cluster Toolkit. Thanks go out to Phil Papadopoulos at NPACI for making the Maui RPM generally available! (See the Rocks Homepage).
||Maui to be included in the OSCAR (Open Source Cluster Applications Resources) distribution containing "the best known methods for building, programming, and using clusters" (See OSCAR Homepage).