MigShm Cache consistancy
Jason Coutu <[email protected]>
| Newsgroups | gmane.linux.cluster.openmosix.devel |
|---|---|
| Message-ID | <[email protected]> |
I've been beating my head around several issues with MigShm recently. At first I couldn't get it to work, even with rpm installs of patched kernels. Eventually I got it to 'work' with the stress test that is provided on the MigShm home page. And I thought I was home free, untill I noticed that the semaphores in the stress test were all backwards, none of them would halt a process when there was not enough resources because the ordering was backwards. Until someone Is willing to step up and maintain MigShm (including fixing flushing, dirty cache, and semaphore issues) I must recomend that it be dropped from any plans to include it in the next openMosix release. The few applications that have benifited from MigShm do not use semaphore locking or need tight cache consistancy. Jason Coutu