Jaeyoung Yoon Computer Sciences Department University of Wisconsin-Madison Virtual Machines in Condor.

Slides:



Advertisements
Similar presentations
Remus: High Availability via Asynchronous Virtual Machine Replication
Advertisements

Virtual Machine Technology Dr. Gregor von Laszewski Dr. Lizhe Wang.
Condor Project Computer Sciences Department University of Wisconsin-Madison Condor's Use of the Cisco Unified Computing System.
PlanetLab Operating System support* *a work in progress.
Virtual Machines What Why How Powerpoint?. What is a Virtual Machine? A Piece of software that emulates hardware.  Might emulate the I/O devices  Might.
Condor and GridShell How to Execute 1 Million Jobs on the Teragrid Jeffrey P. Gardner - PSC Edward Walker - TACC Miron Livney - U. Wisconsin Todd Tannenbaum.
Efficiently Sharing Common Data HTCondor Week 2015 Zach Miller Center for High Throughput Computing Department of Computer Sciences.
Virtualization for Cloud Computing
The Origin of the VM/370 Time-sharing system Presented by Niranjan Soundararajan.
1 MASTERING (VIRTUAL) NETWORKS A Case Study of Virtualizing Internet Lab Avin Chen Borokhovich Michael Goldfeld Arik.
Jaeyoung Yoon Computer Sciences Department University of Wisconsin-Madison Virtual Machine Universe in.
1 Virtualization and Virtual Machines Sarah Diesburg 1/10/2013 COP 5641.
Derek Wright Computer Sciences Department, UW-Madison Lawrence Berkeley National Labs (LBNL)
Zach Miller Condor Project Computer Sciences Department University of Wisconsin-Madison Flexible Data Placement Mechanisms in Condor.
To run the program: To run the program: You need the OS: You need the OS:
Using Virtualization in the Classroom. Using Virtualization in the Classroom Session Objectives Define virtualization Compare major virtualization programs.
Condor Project Computer Sciences Department University of Wisconsin-Madison Virtual Machines in Condor.
Jaime Frey Computer Sciences Department University of Wisconsin-Madison Virtual Machines in Condor.
Design and Implementation of a Single System Image Operating System for High Performance Computing on Clusters Christine MORIN PARIS project-team, IRISA/INRIA.
Cheap cycles from the desktop to the dedicated cluster: combining opportunistic and dedicated scheduling with Condor Derek Wright Computer Sciences Department.
Tanenbaum 8.3 See references
Zach Miller Computer Sciences Department University of Wisconsin-Madison What’s New in Condor.
Miron Livny Computer Sciences Department University of Wisconsin-Madison Harnessing the Capacity of Computational.
Alain Roy Computer Sciences Department University of Wisconsin-Madison An Introduction To Condor International.
Distributed Systems Early Examples. Projects NOW – a Network Of Workstations University of California, Berkely Terminated about 1997 after demonstrating.
Yavor Todorov. Introduction How it works OS level checkpointing Application level checkpointing CPR for parallel programing CPR functionality References.
Hands-On Virtual Computing
The Glidein Service Gideon Juve What are glideins? A technique for creating temporary, user- controlled Condor pools using resources from.
Copyright © 2011 EMC Corporation. All Rights Reserved. MODULE – 6 VIRTUALIZED DATA CENTER – DESKTOP AND APPLICATION 1.
Using Virtualization in the Classroom. Using Virtualization in the Classroom Session Objectives Define virtualization Compare major virtualization programs.
Virtualization Concepts Presented by: Mariano Diaz.
Benefits: Increased server utilization Reduced IT TCO Improved IT agility.
Hao Wang Computer Sciences Department University of Wisconsin-Madison Security in Condor.
1 1 Vulnerability Assessment of Grid Software Jim Kupsch Associate Researcher, Dept. of Computer Sciences University of Wisconsin-Madison Condor Week 2006.
Grid Computing I CONDOR.
Greg Thain Computer Sciences Department University of Wisconsin-Madison cs.wisc.edu Interactive MPI on Demand.
ETICS All Hands meeting Bologna, October 23-25, 2006 NMI and Condor: Status + Future Plans Andy PAVLO Peter COUVARES Becky GIETZEL.
Condor: High-throughput Computing From Clusters to Grid Computing P. Kacsuk – M. Livny MTA SYTAKI – Univ. of Wisconsin-Madison
The Roadmap to New Releases Derek Wright Computer Sciences Department University of Wisconsin-Madison
Derek Wright Computer Sciences Department University of Wisconsin-Madison MPI Scheduling in Condor: An.
Todd Tannenbaum Computer Sciences Department University of Wisconsin-Madison Quill / Quill++ Tutorial.
1 Condor BirdBath SOAP Interface to Condor Charaka Goonatilake Department of Computer Science University College London
Derek Wright Computer Sciences Department University of Wisconsin-Madison Condor and MPI Paradyn/Condor.
Derek Wright Computer Sciences Department University of Wisconsin-Madison New Ways to Fetch Work The new hook infrastructure in Condor.
Security Vulnerabilities in A Virtual Environment
Peter Couvares Associate Researcher, Condor Team Computer Sciences Department University of Wisconsin-Madison
Greg Thain Computer Sciences Department University of Wisconsin-Madison Configuring Quill Condor Week.
Docker and Container Technology
Dan Bradley Condor Project CS and Physics Departments University of Wisconsin-Madison CCB The Condor Connection Broker.
Matthew Farrellee Computer Sciences Department University of Wisconsin-Madison Condor and Web Services.
Todd Tannenbaum Computer Sciences Department University of Wisconsin-Madison Condor NT Condor ported.
Jaime Frey Computer Sciences Department University of Wisconsin-Madison Condor and Virtual Machines.
Unit 2 VIRTUALISATION. Unit 2 - Syllabus Basics of Virtualization Types of Virtualization Implementation Levels of Virtualization Virtualization Structures.
Gabi Kliot Computer Sciences Department Technion – Israel Institute of Technology Adding High Availability to Condor Central Manager Adding High Availability.
VIRTUAL MACHINE – VMWARE. VIRTUAL MACHINE (VM) What is a VM? – A virtual machine (VM) is a software implementation of a computing environment in which.
Condor on Dedicated Clusters Peter Couvares and Derek Wright Computer Sciences Department University of Wisconsin-Madison
Workspace Management Services Kate Keahey Argonne National Laboratory.
Intro To Virtualization Mohammed Morsi
Greg Quinn Computer Sciences Department University of Wisconsin-Madison Privilege Separation in Condor.
Virtualization for Cloud Computing
HTCondor Networking Concepts
HTCondor Networking Concepts
HTCondor Security Basics
Operating System Structure
Building Grids with Condor
Introduction to Operating Systems
OS Virtualization.
Basic Grid Projects – Condor (Part I)
Condor-G Making Condor Grid Enabled
Presentation transcript:

Jaeyoung Yoon Computer Sciences Department University of Wisconsin-Madison Virtual Machines in Condor

Virtual Machines › VMware  Pros: Full virtualization, no modification for guest OS, and checkpoint/restart capability  Cons: Commercial product and performance issue compared to Xen › Xen  Pros: Open source, good performance, checkpoint/restart and live migration capability  Cons: Requires OS modification and must divide memory between host and VMs in advance › UML (User Mode Linux) etc.

Benefit of using Virtual Machine in Condor › Sandbox  Security and Isolation › Independent environment  Customizing environment for Condor › Several OS’s on a single physical machine  Support for a wider variety of jobs › Finer Resource Control  Assign memory size to each VM explicitly › Checkpoint and migration  All memory of VM can be saved(or suspended) and restarted(or resumed) later

Difficulty of using Virtual Machine in Condor › Hard to manage system memory efficiently › Need to know some information of host machine inside VM › Need to setup some environments in VM › If a VM cannot use the distributed file system, Condor’s file transfer or remote IO mechanism should be used › Need IP address for each VM

How to use VM in Condor Scenario 1 › Already launched VM is ready to be used as a execution machine for Condor jobs › Condor daemons should be installed and run on both the virtual and the host machine, which are all exposed to the pool › Condor startd on the host machine controls when a launched VM is used for Condor. › Supported by Condor and all future releases › Pros: Easy to implement › Cons: Inefficient memory management

Scenario 1 Host Machine Virtual machine Central Manager Collector Startd Negotiator Startd Execution machine Schedd Submit machine Communication pathway

Current Implementation How can VM get the information for host machine 1. Query ClassAd for host ClassAd for VM 2. ClassAd for host machine Virtual Machine VMP_HOST_MACHINE = host.domain.com ClassAd for host Name = host.domain.com” TotalLoadAvg = KeyboardIdle = 50 …. Host Machine VMP_VM_LIST = vmware1.domain.com Name = “vmware1.domain.com” TotalLoadAvg = KeyboardIdle = …. Before Query Name = “vmware1.domain.com” TotalLoadAvg = KeyboardIdle = … HOST_Name=“host.domain.com” HOST_TotalLoadAvg = HOST_KeyboardIdle = 50 … After Query

Current Implementation How does a VM get permission from host machine 1. Send VM_REGISTER ClassAd for VM Virtual Machine VMP_HOST_MACHINE = host.domain.com ClassAd for host START = ((KeyboardIdle > 150 ) && ( LoadAvg <= 0.3 )) Host Machine VMP_VM_LIST = vmware1.domain.com START = False If host status == (‘owner’ | ‘unclaimed’) START = ((KeyboardIdle > 150 ) && ( LoadAvg <= 0.3 )) else 2. Reply permission START = ((KeyboardIdle > 150 ) && (HOST_KeyboardIdle > 150 ) && ( LoadAvg <= 0.3 ) && (HOST_TotalLoadAvg <= 0.3 )) If permission == yes else START = False

Issues in current implementation for Scenario 1 › Problem: › Problem: host machine cannot be used for Condor any more after sending a permission to a virtual machine. › Possibility: › Possibility: A user may want to use both virtual and host machine in a SMP machine. › Possible solution: › Possible solution: After sending permission, host machine does not change START expression. Instead, the virtual machine sends its status to host machine periodically and host machine decides the permission for each virtual machine when a Condor job is assigned.

How to use VM in Condor Scenario 2 › Virtual Machine is launched on demand to serve a Condor job › Checkpoint and migration per virtual machine base can be used › Startd on host machine may have to advertise more than one OS. › A specific daemon in virtual machine needs to communicate with host machine. The daemon receives a command from host machine and executes it when a Condor job is assigned.

How to use VM in Condor Scenario 2 › Not yet implemented in Condor but we hope to do it soon. › Pros: Efficient memory management › Cons: Complex to implement

Scenario 2 Host Machine Virtual machine Daemon Startd Execution machine Negotiator Central Manager Collector Schedd Submit machine 1 launching Starter Shadow Communication pathway Creating/forking process daemon Submit machine 2 Virtual machine schedd Host Machine

Issues in Scenario 2 › Stop VM and save the entire memory of VM instead of suspending an executed Condor job when a user returns to the host machine › During migration if there is no shared file system, files used by a Condor job, including a program file, should be transferred because it is very hard to copy entire disk image. › Xen live migration technique can be effectively used for direct migration without checkpointing.

Scenario 2 - Migration Host Machine Virtual machine Central Manager Collector Daemon Negotiator Startd Execution machine 1 Schedd Submit machine Host Machine Virtual machine Daemon Startd Execution machine 2 migration launching Communication pathway before migration Communication pathway after migration Shadow Starter Creating/forking process

› Virtual Machine offers flexible solution in Condor  Sandbox for security  Can provide more than one OS on a single physical machine  Can provide customized environment for Condor › Scenario 1 has been already supported since Condor › Scenario 2 is not yet implemented in Condor. That is a future work in Condor. Summary