Roy Bixler
rcb@press.uchicago.edu
Tue Oct 5 16:15:34 UTC 2004
On Tuesday 05 October 2004 09:07 am, Vernon Schryver wrote: > On the contrary, most modern free UNIX systems are absolutely terrible > in how they handle their disk caches for large files. They generally > do no have unified page and file caches, but prefer to pump pages among > disk files, swap, buffer cache, and VM. They generally cannot be > convinced to swap directly to and from memory mapped files, which is > why many regularly stall and busily pump pages of the DCC database > among RAM, the DCC files, and swap space! Linux is particularly bad > in this area. I used to notice this in using earlier versions of Linux 2.x.y, where x<6, kernels even with ordinary desktop loads. I understand that in the Linux 2.6 series that they have actually implemented a unified buffer cache in the VM system. In each case, where x<6 and y is maximal (i.e. the latest version of that series), the memory management performance is greatly improved over the earlier versions. I don't run a DCC server, but you may want to try upgrading to the latest Linux kernel to see if that helps. I have also tried FreeBSD and that also solved the problem. -- Roy Bixler <rcb@press.uchicago.edu> The University of Chicago Press http://www.press.uchicago.edu -or- http://www.journals.uchicago.edu
More information about the DCC
mailing list