In relation to the amount of swap space you need to configure for a system, what would be the consequences of setting the following kernel parameter to the value displayed?
swapmem_on 0
2:
The buffer-cache is dynamic in nature, i.e., its size can increase as well as decrease depending on system requirements. On our system, the kernel parameters controlling the size of the dynamic buffer cache have their default values. We are performing a significant amount of filesystem IO, increasing the size of the buffer cache beyond its 5 percent initial value. When will the amount of memory allocated to the buffer cache actually fall?
3:
When memory pressure is high, processes can have data paged-out to a swap device. When a process has data paged-out, the process is allocated a swchunk amount of swap space even though the system only needs to page-out as little as one page. Give at least one reason why you think the system allocates an entire swchunk rather than simply allocating a single page from the swap device.
4:
Why does the swapper process no longer swap an entire process to disk? What is the purpose of the swapper process? How do pages from a swapped process get put out on the swap device?
5:
Dump devices need to be contiguous in nature. Why is this so? Would it be a good idea to mirror a dump device?
