Condor NT Condor ported to Win 32 Todd

Condor NT Condor ported to Win 32 Todd Tannenbaum Computer Sciences Department University of Wisconsin-Madison tannenba@cs. wisc. edu http: //www. cs. wisc. edu/condor

Overview › Intro to Condor NT › What does Condor NT do? › How does Condor NT differ from › › Condor for Unix? What are the current limitations of Condor NT? Future Work www. cs. wisc. edu/condor 2

Intro to Condor NT › First pre-release at Condor ver 6. 1. 8 › “Deep port” of Condor › Daemons run as a system Service › under the Local. System account Shares as much source code with Condor for Unix as possible www. cs. wisc. edu/condor 3

Condor NT Downloads www. cs. wisc. edu/condor 4

What can it do? › Almost everything Condor for Unix can… h. Submit, run, manage queues of jobs • Jobs run “in the background” h. Nearly all Condor tools included h. Class. Ads • Full compliment of attributes (load average, RAM, benchmarks, free swap, key/mouse idle times, image size, CPU usage, etc) h. Everything needed for a Central Manager www. cs. wisc. edu/condor 5

What can it do? (cont) › Support for SMP machines › Several security mechanisms (more later…) › Suspend, continue, soft-kill (WM_CLOSE), › › › hard-kill jobs Correctly manage multi-process jobs Send email notifications Yada, yada, … www. cs. wisc. edu/condor 6

What’s missing? › Only VANILLA universe included h. No STANDARD, PVM, GLOBUS, SCHEDULER universe h. Note: MPI being done on both Unix and Win 32 › Ability to run the job as the submitting › › user Ability to access shared volumes as the submitting user So – who does the job run as, how does the job get its files? www. cs. wisc. edu/condor 7

Job Start on Condor NT › On execute machine, Condor creates h. New temporary user account h. New temporary working directory h. New temporary, non-visible desktop › Permissions (ACLs) set › Files transferred by Condor › Job spawned www. cs. wisc. edu/condor 8

While Job is Running… › Condor watches the job and updates dynamic attributes about the job in the job Class. Ad h. Disk usage, cpu usage, … › Enforces the machine owner’s policy www. cs. wisc. edu/condor 9

On Job Vacate/Exit… › Condor conditionally transfers any output files back to the submit machine h. Can be told filenames, or automatically send back files which have changed h. File transfers are atomic › Cleanup www. cs. wisc. edu/condor 10

Some points on shared (network) filesystem access › On Condor Unix, VANILLA requires a shared filesystem h. Not true on Condor NT › Condor NT can access a shared filesystem h… but only as user “Guest” or only if the share password is provided by the job www. cs. wisc. edu/condor 11

Difficulties of running as the user › Forwarding credentials problem h. Windows NTLM in NT 4. 0 can impersonate the peer on a socket, but only one “jump” A B C › On Windows NT, cannot just setuid() www. cs. wisc. edu/condor 12

Current Work To Do › Improve situation for access to shared filesystem h. As user “condor”, or h. As user who submitted the job › Run jobs as the submitting user h. On NT 4. 0 : store the password, forward it encrypted h. On Windows 2000: same or PKI www. cs. wisc. edu/condor 13

Current Work Todo, Cont. › Windows 2000 support h. Current release mostly works on Win 2 k… h. Take advantage of Win 2 k enhancements › Add in Scheduler Universe h. And therefore DAGMan support › Add in the MPI Universe www. cs. wisc. edu/condor 14

Future Work › Add remaining missing Condor Universes h. STANDARD • Requires addition of process checkpoint and/or remote system call h. GLOBUS • Requires Globus Toolkit client libs on Win 32 h. PVM www. cs. wisc. edu/condor 15

Thank You! www. cs. wisc. edu/condor 16