blob: 742523e145bc4329f31c51712f2fe0099b5ae0ba (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
|
Warning:
The xenpaging code is new and not fully debugged.
Usage of xenpaging can crash Xen or cause severe data corruption in the
guest memory and its filesystems!
Description:
xenpaging writes memory pages of a given guest to a file and moves the
pages back to the pool of available memory. Once the guests wants to
access the paged-out memory, the page is read from disk and placed into
memory. This allows the sum of all running guests to use more memory
than physically available on the host.
Requirements:
xenpaging relies on Intel EPT or AMD RVI, other hardware is not
supported. Only HVM guests are supported. The dom0 kernel needs
paging-aware backend drivers to handle paged granttable entries.
Currently only dom0 kernels based on classic Xen Linux support this
functionality.
Usage:
Up to now xenpaging is not integrated into libxl/xend, so it has to be
started manually for each guest.
Once the guest is running, run xenpaging with the guest_id and the path
to the pagefile:
/usr/lib/xen/bin/xenpaging -f /path/to/page_file -d dom_id &
Once xenpaging runs it needs a memory target, which is the memory
footprint of the guest. This value (in KiB) must be written manually to
xenstore. The following example sets the target to 512MB:
xenstore-write /local/domain/<dom_id>/memory/target-tot_pages $((1024*512))
Now xenpaging tries to page-out as many pages to keep the overall memory
footprint of the guest at 512MB.
Todo:
- integrate xenpaging into libxl
# vim: tw=72
|
