KernelTrap: Measuring The Effect Of Swappiness

"In a recent report by Ray Bryant generated on a 32 CPU Altix system, it was suggested that tuning how much a system swaps with /proc/sys/vm/swappiness has a different effect depending on the version of the 2.6 Linux kernel that's being used. Andrew Morton acknowledged that this could be the accidental side-effects from changes made in the page reclaim code. Going forwarded he noted that he intends to minimize changes to this behavior:

"'Hopefully very little. Unless we choose to deliberately change the swapout behaviour. The code in there is complex and as you've seen, has surprising interactions. And changes have been made without sufficiently broad testing. So I'll be setting the bar much higher for changes to vmscan.c. It takes a *lot* of work to demonstrate that a change in there does what it's supposed to do without breaking other things...'"

Complete Story

Related Stories:

• Linux Magazine: Programming Linux 2.6(Sep 05, 2004)
• KernelTrap: Linux: 2.6.7-ck5, Reviewing Autoregulated Swappiness(Jul 11, 2004)
• Kernel Trap: Linux: Is Swap Necessary?(Jun 02, 2004)
• KernelTrap: VM Swappiness Autoregulation(Oct 27, 2003)