Re: Using Mosix: How to do 100 simulation runs, 7 at a time
Tony Travis <[email protected]>
| Newsgroups | gmane.linux.cluster.openmosix.general |
|---|---|
| Message-ID | <[email protected]> |
Matt W wrote: > > Greetings, > > Im trying to help out a user to kick of a long running job of 100 > compute-intensive runs, but only want to run no more than 7 > of them at any given time on the cluster. > > Can someone comment on a good way to accomplish this on the mosix cluster? Hello, Matt. You could do it with a 'semaphore', which is one method of controlling access to a limited resource - You would need a wrapper to check the semaphore and start jobs. When a job runs, the wrapper increments the semaphore and you don't permit more than 7 instances to start. When a job ends, the wrapper should decrement the semaphore - any jobs that are blocked waiting for the semaphore will run when they poll the semaphore and get a value of less than 7. > Sample code would be great. Yes, it would but I don't have any ;-) Seriously, though I'm interested in doing something similar myself and I've got as far as thinking about how to do it without using a DRM (Distributed Resource Manager) like SGE (Sun Grid Engine) or similar. The strategy I've outlined is how access to limited resources is done at the kernel level... > BTW: We have 20 node Debian Sarge OpenMosix Cluster of XEON > and OPTERON nodes running the 2.4.26 SMP kernel + libATA patch We have a 64-node Athlon cluster with Opteron servers running Red Hat 9 with an openMosix kernel, but I'm in the process of migrating to Debian. Best wishes, Tony. -- Dr. A.J.Travis, | mailto:[email protected] Rowett Research Institute, | http://www.rri.sari.ac.uk/~ajt Greenburn Road, Bucksburn, | phone:+44 (0)1224 712751 Aberdeen AB21 9SB, Scotland, UK. | fax:+44 (0)1224 716687 ------------------------------------------------------- This SF.Net email is sponsored by xPML, a groundbreaking scripting language that extends applications into web and mobile media. Attend the live webcast and join the prime developer group breaking into this new coding territory! http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642