[ale] Kernel 2.6 : worst-case scenarios if swap full

[Jerry Yu]

On a db server with 16G ram and 1G swap, swap is nearly full. (It
swang between 88 and 96% last week). What's the worst-case scenarios
in terms of performance and/or availabilitu, when kernel decides to
page out a few extra idle pages (more than the free space in swap)
from physical ram?

Assume the following,
* Free physical ram is at 5G, or 9G including buffer/cache.
* no new program will be launched, except stock RHEL 5.1 cron jobs and
a few ssh sessions
* OLTP transaction load against the db may spike 5 times over the
weekly average, with more client connections to db server
* swappiness is at default
* db is the only non-os process running. Mem for db is preallocated.
No tweaks is done to prevent db pages from being paged out