A B (JIRA) wrote:
Interesting observation about the disk cache--thanks for bringing that up. Is it feasible to turn disk cache on for one of the machines that's failing and run the tests for a couple of days to see if that makes a difference? Thanks Øystein...
I can turn the disk cache on for one of the machines. Note that the tinderbox test runs on an supposedly identical machine with the disk cache on, and the failure has never been seen there.
-- Øystein
