From: cira <ma...@dr...> - 2012-06-29 14:09:16
|
On 6/28/2012 05:26, Trent Davis wrote: > Hello, > > I have posted a bug at: > http://sourceforge.net/tracker/index.php?func=detail&aid=3536755&group_id=8642&atid=108642 > > However I'm not sure anyone has had a chance to look at it. > > I have a repeatable test case and willing to help any way I can before > the release becomes stable. > > I have worked around the issue for my personal use by just adding a few > GB at a time to the TM backup (by excluding lots of stuff, then slowly > adding folders back in). > > Should I be posting here or to the -devel list? > > Cheers. Um... before diving into the code I notice that in your bug report that you mention it takes a varying length of data transfer (multi gigabytes) before the signal 11 is triggered. SIGSEGV is caused by an invalid memory reference. How confident are you of your hardware -- client and server? My personal experience has been that a /random/ SIGSEGV occurring under heavy load -- multigigs is a heavy load -- means hardware failure, usually RAM. See: http://www.bitwizard.nl/sig11/ Does the server in question pass the 'compile the linux kernel 100 times' test? Or I guess in this case, compile the solaris kernel 100 times? -cira |