© 2002 Barton P. MillerMarch 4, 2001Tool Dæmon Protocol The Tool Dæmon Protocol: Using Monitoring Tools on Remote Applications Barton P. Miller

Slides:



Advertisements
Similar presentations
MPI Message Queue Debugging Interface Chris Gottbrath Director, Product Management.
Advertisements

Message Queue Dumping MPI Side Document. History MPIR and MQD were designed around 1995 MPIR: Process discovery Formally described in the MPIR document.
Chap 2 System Structures.
Setting up of condor scheduler on computing cluster Raman Sehgal NPD-BARC.
USING THE GLOBUS TOOLKIT This summary by: Asad Samar / CALTECH/CMS Ben Segal / CERN-IT FULL INFO AT:
A Grid Parallel Application Framework Jeremy Villalobos PhD student Department of Computer Science University of North Carolina Charlotte.
Parallel/Concurrent Programming on the SGI Altix Conley Read January 25, 2007 UC Riverside, Department of Computer Science.
RPC Project Using either sockets or TLI, implement Remote Procedure Calls between two distinct machines that are communicating over an Ethernet network.
1 Reminding - ARP Two machines on a given network can communicate only if they know each other’s physical network address ARP (Address Resolution Protocol)
Protection of Agent Teamwork By Jeremy Hall. Agent Teamwork Overview ● Mobile agent framework  AgentTeamwork 2 is a mobile-agent based middleware system.
Smart Home Current Progress Summary. Main Processor – Stellaris.
Web Proxy Server Anagh Pathak Jesus Cervantes Henry Tjhen Luis Luna.
By Justin Thompson. What is SOAP? Originally stood for Simple Object Access Protocol Created by vendors from Microsoft, Lotus, IBM, and others Protocol.
Jaeyoung Yoon Computer Sciences Department University of Wisconsin-Madison Virtual Machines in Condor.
The SAM-Grid Fabric Services Gabriele Garzoglio (for the SAM-Grid team) Computing Division Fermilab.
Miron Livny Computer Sciences Department University of Wisconsin-Madison Harnessing the Capacity of Computational.
Track 1: Cluster and Grid Computing NBCR Summer Institute Session 2.2: Cluster and Grid Computing: Case studies Condor introduction August 9, 2006 Nadya.
Prof. Heon Y. Yeom Distributed Computing Systems Lab. Seoul National University FT-MPICH : Providing fault tolerance for MPI parallel applications.
Tools and Utilities for parallel and serial codes in ENEA-GRID environment CRESCO Project: Salvatore Raia SubProject I.2 C.R. ENEA-Portici. 11/12/2007.
Objectives Configure routing in Windows Server 2008 Configure Routing and Remote Access Services in Windows Server 2008 Network Address Translation 1.
Interrupts and DMA CSCI The Role of the Operating System in Performing I/O Two main jobs of a computer are: –Processing –Performing I/O manage and.
A Metadata Based Approach For Supporting Subsetting Queries Over Parallel HDF5 Datasets Vignesh Santhanagopalan Graduate Student Department Of CSE.
Lecture 15 Introduction to Web Services Web Service Applications.
Grids and Portals for VLAB Marlon Pierce Community Grids Lab Indiana University.
Process Management Working Group Process Management “Meatball” Dallas November 28, 2001.
Support for Debugging Automatically Parallelized Programs Robert Hood Gabriele Jost CSC/MRJ Technology Solutions NASA.
Grid Resource Allocation and Management (GRAM) Execution management Execution management –Deployment, scheduling and monitoring Community Scheduler Framework.
Grid Computing I CONDOR.
COMP3019 Coursework: Introduction to GridSAM Steve Crouch School of Electronics and Computer Science.
DataGrid WP1 Massimo Sgaravatto INFN Padova. WP1 (Grid Workload Management) Objective of the first DataGrid workpackage is (according to the project "Technical.
The Globus Project: A Status Report Ian Foster Carl Kesselman
Data Issues & Recommendations. Data Issues AON meeting last week—18 AON PIs, plus collaborating groups & programs –CADIS, an AON project for cooperative.
CHEP03 Mar 25Mary Thompson Fine-grained Authorization for Job and Resource Management using Akenti and Globus Mary Thompson LBL,Kate Keahey ANL, Sam Lang.
Climate-Weather modeling studies Using a Prototype Global Cloud-System Resolving Model Zhi Liang (GFDL/DRC)
Derek Wright Computer Sciences Department University of Wisconsin-Madison MPI Scheduling in Condor: An.
The Grid the united computing power Jian He Amit Karnik.
Portable Parallel Performance Tools Shirley Browne, UTK Clay Breshears, CEWES MSRC Jan 27-28, 1998.
1 Condor BirdBath SOAP Interface to Condor Charaka Goonatilake Department of Computer Science University College London
PMI: A Scalable Process- Management Interface for Extreme-Scale Systems Pavan Balaji, Darius Buntinas, David Goodell, William Gropp, Jayesh Krishna, Ewing.
Globus Toolkit Massimo Sgaravatto INFN Padova. Massimo Sgaravatto Introduction Grid Services: LHC regional centres need distributed computing Analyze.
Network Address Translation External/ Internal/. OVERLOADING In Overloading, each computer on the private network is translated to the same IP address;
© 2001 Barton P. MillerParadyn/Condor Week (12 March 2001, Madison/WI) The Paradyn Port Report Barton P. Miller Computer Sciences Department.
1 MPI_Connect and Parallel I/O for Distributed Applications Dr Graham E Fagg Innovative Computing Laboratory University of Tennessee Knoxville, TN
Derek Wright Computer Sciences Department University of Wisconsin-Madison Condor and MPI Paradyn/Condor.
How to for compiling and running MPI Programs. Prepared by Kiriti Venkat.
PTools Annual Meeting, Knoxville, TN, September 2002 The Tool Daemon Protocol: Defining the Interface Between Tools and Process Management Systems.
TeraGrid Gateway User Concept – Supporting Users V. E. Lynch, M. L. Chen, J. W. Cobb, J. A. Kohl, S. D. Miller, S. S. Vazhkudai Oak Ridge National Laboratory.
© 2001 B. P. Miller & M. Livny (12-14 March 2001)Paradyn/Condor Week Agenda Paradyn/Condor Week 2001 Barton P. Miller Miron Livny
Globus and PlanetLab Resource Management Solutions Compared M. Ripeanu, M. Bowman, J. Chase, I. Foster, M. Milenkovic Presented by Dionysis Logothetis.
Globus Grid Tutorial Part 2: Running Programs Across Multiple Resources.
2/22/2001Greenbook 2001/OASCR1 Greenbook/OASCR Activities Focus on technology to enable SCIENCE to be conducted, i.e. Software tools Software libraries.
Condor Services for the Global Grid: Interoperability between OGSA and Condor Clovis Chapman 1, Paul Wilson 2, Todd Tannenbaum 3, Matthew Farrellee 3,
Miron Livny Computer Sciences Department University of Wisconsin-Madison Condor and (the) Grid (one of.
© 2001 Week (14 March 2001)Paradyn & Dyninst Demonstrations Paradyn & Dyninst Demos Barton P. Miller Computer.
Matthew Farrellee Computer Sciences Department University of Wisconsin-Madison Condor and Web Services.
1 Advanced MPI William D. Gropp Rusty Lusk and Rajeev Thakur Mathematics and Computer Science Division Argonne National Laboratory.
Grid Execution Management for Legacy Code Architecture Exposing legacy applications as Grid services: the GEMLCA approach Centre.
HTCondor’s Grid Universe Jaime Frey Center for High Throughput Computing Department of Computer Sciences University of Wisconsin-Madison.
Tutorial on Science Gateways, Roma, Catania Science Gateway Framework Motivations, architecture, features Riccardo Rotondo.
POE Parallel Operating Environment. Cliff Montgomery.
Emulating Volunteer Computing Scheduling Policies Dr. David P. Anderson University of California, Berkeley May 20, 2011.
The Distributed Application Debugger (DAD)
HTCondor Networking Concepts
HTCondor Networking Concepts
Peter Kacsuk – Sipos Gergely MTA SZTAKI
CompSci 725 Presentation by Siu Cho Jun, William.
Introducing To Networking
Building Grids with Condor
Initial job submission and monitoring efforts with JClarens
Parallel I/O for Distributed Applications (MPI-Conn-IO)
Presentation transcript:

© 2002 Barton P. MillerMarch 4, 2001Tool Dæmon Protocol The Tool Dæmon Protocol: Using Monitoring Tools on Remote Applications Barton P. Miller Computer Sciences Department University of Wisconsin Madison, Wisconsin USA

– 2 –© 2002 Barton P. Miller Tool Daemon Protocol The Challenge Consider remote process management environments: qCondor qMPI Portable MPI (such as MPICH) Vendor provided MPI (such as IBM, Compaq, Sun) qGlobus Each of these environments needs to monitor and control the state of its application processes.

– 3 –© 2002 Barton P. Miller Tool Daemon Protocol Typical Process Manager Process manger: Starts the remote job Monitors its status Controls the job Sets up file I/O Sets up standard I/O Remote Process Manager monitor/ control Application Process Application Process Remote Host

– 4 –© 2002 Barton P. Miller Tool Daemon Protocol Typical Process Manager Remote Process Manager monitor/ control Application Process Application Process Remote Host Tool Dæmon Process ? ? The run-time tool? Also may want to start process (or attach to it) Also needs to monitors its status Also may want to control the job Needs to communicate with its front-end.

– 5 –© 2002 Barton P. Miller Tool Daemon Protocol Typical Process Manager Remote Process Manager monitor/ control Application Process Application Process Remote Host ? ? Tool Dæmon Process So, who wins?

– 6 –© 2002 Barton P. Miller Tool Daemon Protocol Typical Process Manager Remote Process Manager monitor/ control Application Process Application Process Remote Host Local Host ? ? Tool Front-End Process Tool Dæmon Process

– 7 –© 2002 Barton P. Miller Tool Daemon Protocol Current State of Affairs qEach process manager starts and controls processes in its own way. E.g., even within MPI: IBM POE MPI, SGI Origin MPI and MPICH all work differently. MPI has no standard process control! qSpecialized cases of a specific tool working with a specific environment e.g., TotalView debugger working with MPICH. qThe result is an m  n combination of m process managers and n tools. Bottom line: need a standard interface for process managers and tools to coexist: the Tool Dæmon Protocol (TDP).

– 8 –© 2002 Barton P. Miller Tool Daemon Protocol The Basic TDP Steps 1.Create, but don’t start, new application process. 2.If necessary, create tool daemon process. 3.Pass basic information to tool daemon: e.g., 1.Application PID. 2.Front-end host/port number. 3.Standard I/O host/port number. 4.Tool daemon processes application: 1.For a debugger, read symbols 2.For Paradyn/dyninst, parse the executable. 5.Start the application process 6.Respond to changes in the application state. 7.Respond to changes in the tool daemon’s state.

– 9 –© 2002 Barton P. Miller Tool Daemon Protocol Step 1a: Request Process Creation Remote Host Local Host Tool Front-End Process Remote Process Manager Create process

– 10 –© 2002 Barton P. Miller Tool Daemon Protocol Step 1b: Create Application Process Application Process Remote Host Local Host Tool Front-End Process Remote Process Manager

– 11 –© 2002 Barton P. Miller Tool Daemon Protocol Step 2: Create Tool Dæmon Application Process Remote Host Local Host Tool Dæmon Process Remote Process Manager Tool Front-End Process

– 12 –© 2002 Barton P. Miller Tool Daemon Protocol Step 3: Pass Info to Tool Dæmon Application Process Remote Host Local Host Remote Process Manager Tool Front-End Process PID, host/port pairs Tool Dæmon Process

– 13 –© 2002 Barton P. Miller Tool Daemon Protocol Step 4: Process the Application Remote Host Local Host Remote Process Manager Tool Front-End Process Tool Dæmon Process Application Process

– 14 –© 2002 Barton P. Miller Tool Daemon Protocol Step 5: Start the Application Remote Host Local Host Remote Process Manager Tool Front-End Process Tool Dæmon Process Application Process

– 15 –© 2002 Barton P. Miller Tool Daemon Protocol Interesting Challenges 1.Return codes and state changes? 2.Who control standard input and output? 3.Private networks and firewalls: 1.Tool dæmon communicating to front-end. 2.Application process sending standard I/O.

– 16 –© 2002 Barton P. Miller Tool Daemon Protocol Challenge: Firewalls and Private Nets Remote Host Local Host Remote Process Manager Tool Front-End Process Tool Dæmon Process Application Process Firewall X

– 17 –© 2002 Barton P. Miller Tool Daemon Protocol Challenge: Firewalls and Private Nets Remote Host Local Host Remote Process Manager Tool Front-End Process Tool Dæmon Process Application Process Firewall Comm Proxy

– 18 –© 2002 Barton P. Miller Tool Daemon Protocol Challenge: Firewalls and Private Nets qWhen tool daemon is started, pass in the host/port number of its front-end process. qIf there is a communication proxy, then: Tool daemon will receive host/port of the proxy, so daemon connects to proxy. Proxy will connect to the tool front-end, mapping the host/port (similar to NAT). qApplication connecting to console for standard I/O works the same way.

– 19 –© 2002 Barton P. Miller Tool Daemon Protocol The Condor/Paradyn Scenario Condor Starter monitor/ control Application Process Application Process Remote Host Local Host Paradyn Front-End Paradyn Dæmon

– 20 –© 2002 Barton P. Miller Tool Daemon Protocol The Path Forward qProduce a prototype implementation to expose technical challenges: Paradyn running under Condor Ana Cortes and Miquel Senar (UAB/UW) A presentation tomorrow about the first steps. qGoal is to produce a standard set of libraries for process managers and tool daemons. qInvolve a wider community in this standards effort Initially: ANL (Gropp and Lusk), Etnus (Cownie and Delsignore), Globus (Kesselman)

– 21 –© 2002 Barton P. Miller Tool Daemon Protocol The Path Forward qThis is just the beginning of an important effort qYour input and participation is welcome and encouraged. This project is supported by the hard work and advice of the Condor and Paradyn teams!