Hello!
Here is the original question:
Harinder Singh writes:
> Has anyone heard of lazy swapping?
> What are the benefits of lazy swapping?
> How do we implement it?
Remove the symbolic link from /sbin/swapdefault to your primary swap
partition, for example:
SOLUTION 1:
A set of the same processes under the same conditions will typically use
much less swap space under lazy swapping than under conservative
swapping, since under lazy swapping swap space is only allocated for
those pages paged out of the process, rather than for the entire process
address space. The only risk is that under lazy swapping a process may
be created and run for a while, then be unable to swap because of
insufficient swap space, and be killed by the kernel. With conservative
swapping the swap space the process would need is allocated at the time
the process starts, so that if there is insufficient swap space it is
killed very early in its lifetime.
mv /sbin/swapdefault /sbin/swapdefault.disabled
SOLUTION 2:
In "normal" swapping space is allocated in the swap area upon process
creation. If there is not enough space available the process creation
will fail. In "lazy" swapping they swap area is not checked swapping
needs to be done. If there is no space available the process terminates
Now under Digital Unix the swappin mode is controled by the file
/sbin/swapdefault, if the file is not there lazy mode is used.
- Eric Crane
Westinghouse Electric
Thanks to everyone for the response!!
Harry
Received on Fri May 16 1997 - 00:30:08 NZST