> With DCSS, will it be first fit, or moving cursor allocation? First fit is > optimal, vdisk is moving cursor. DCSS > should be paged out with the same rules as vdisk, but this needs to be > verified in real life. The CPU savings are > "obvious", but are we talking 1 percent of 1 percent of an IFL? Swapping at > 1000 per second to vdisk used about 10% of > an 890 CPU for reference. That > is a pretty sizable target on an 890, not so sizeable on a z10, and nobody > swaps at > 1000 per second regularly. most installations swap even across 100 servers a > total of less than 100 per second, with > the average generally being less than 1. > So if swap rate is high, then probably there is a configuration problem, and > if swapping is low, what is the gain?
I smell a re-validation of the claim "VDISK for Linux swap is the best solution" on Velocity's website coming up.... ;-) Pieter Brabant Water N.V. Postbus 1068 5200 BC 's-Hertogenbosch http://www.brabantwater.nl Handelsregister: 16005077 ---------------------------------------------------------------------- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
