|
From: Eyal L. <ey...@ey...> - 2004-02-10 08:44:30
|
Nicholas Nethercote wrote: > > On Sat, 7 Feb 2004, Eyal Lebedinsky wrote: > > > A program (sched) is execing another (init) which then > > spawns (fork/exec) another. The failure is always in > > this last execvp() before the target executable starts. > > Any chance you can reduce it to a small test case? I doubt it. I already tried once and failed. This seems to happen in a very irregular way, and rarely. Nevertheless, If I manage to get anything like reproducability then I will attempt to reduce the system to something that I can submit as a test case. I will keep trying though. Is there an option to induce vg to generate a more detailed report, which may give a clue to an educated developer? -- Eyal Lebedinsky (ey...@ey...) <http://samba.org/eyal/> |