On Nov 16, 2009, at 11:53 AM, Andrew Klaassen wrote:

--- On Mon, 11/16/09, Michael Oliver mcoliver@gmail.com wrote:

Recently a few Maya scenes coming through the render farm are failing out with SIGTERM 11 errors on our CentOS 5.3 x64 render boxes.

Signal 11 is most often (though not always) a sign of bad physical
memory. If the errors are happening consistently on certain boxes,
I'd try trading memory modules between machines to see if the
failures follow the memory modules.

Andrew

heh. I can assure you that as a developer I see thousands of software-induced SIGSEGVs for every hardware instance. In these cases it is usually due to a bug of some sort. I have scars to prove it! Frequent suspects are:

• dereference null pointer
• walk off the end of an array
• accessing previously freed memory
• not heading malloc failure
• threading race condition stomping on "live" data

obviously a large code base like maya is not immune to bugs, but you might start with removing any recently added plug-ins to see if the problem is actually in a plug-in rather than in maya itself?

• David

---

StudioSysAdmins-Discuss mailing list StudioSysAdmins-Discuss@mailman.studiosysadmins.com http://mailman.studiosysadmins.com/mailman/listinfo/studiosysadmins-discuss