On 1/19/2011 12:07 AM, William Kenworthy wrote:
Do you have a verifiable (as in from a knowledgeable source) reference
for this? - it goes against a lot of what I found googling a year ago
where swap size was dependent on CPU architecture (i.e.,
zeon/opteron/athlon etc), not 32/64bit.)
You know the more I look into this the weirder it gets.
Number of swap devices with later kernels included
http://www.kernel.org/doc/man-pages/online/pages/man2/swapon.2.html#NOTES
However because the man page mkswap is waaaaaay out of date I'm not
inclined to trust swapon's man page either.
Starting in 2003 we see that mkswap actually had the 2GB limit whereas
the kernel already had much higher limits.
http://lkml.indiana.edu/hypermail/linux/kernel/0306.0/1725.html
The same character revist the issues two years later in 2005.
http://lkml.indiana.edu/hypermail/linux/kernel/0506.0/0136.html
As far as I can tell it comes down to cluster size, bitness of your OS,
and amount of RAM you're willing to dedicate to managing swap.
kashani
