On Sun, Sep 08, 2002 at 06:29:33PM -0700, Trevor Marshall wrote: > Jw, > Respectfully, > My experience is to shy away from any piece of software which the > developers feel is inviolate > > Although often the hardware and kernel certainly could be at fault it is > wrong to assume they are, and such a response usually indicates that I am > wasting my time reporting this problem. >
I'm sorry if i gave that impression. Neither i nor anyone else here feel that rsync is inviolate. It does have plenty of areas for improvement. Rsync does little more than malloc, fork, exec, stat, read, write, unlink and sometimes a little networking. In other words, basic I/O. Unless you try to rsync /proc it doesn't do anything that could be that dangerous. A _really_ big tree could cause OOM but few sites are likely to go there. What you have described is that in a string of system crashes you have managed to get one incident where rsync has reported an error without the system crashing. By definition the kernel is immune to application errors. If you find an app that crashes the kernel you have found a test case for a kernel bug. If rsync is doing something wrong that incidentally causes the kernel to bomb, OK we want to fix it. But let's isolate the problem. If the kernel is crashing we can't trust application behavior. -- To unsubscribe or change options: http://lists.samba.org/mailman/listinfo/rsync Before posting, read: http://www.tuxedo.org/~esr/faqs/smart-questions.html
