IO Limiting (Priorities) [message #35737] |
Sun, 19 April 2009 18:46 |
CodyRo
Messages: 1 Registered: April 2009 Location: Systems Administrator
|
Junior Member |
|
|
I've searched over the forums and don't think this is clarified anywhere, though I could be mistaken (and if I am, excuse me ).
In the Wiki IO priority is described to work as the following:
Quote: |
Priority range is 0-7. The more priority a container has, the more time for using block devices this container will obtain.
|
This seems fairly clear - the higher the priority given, the more IO the VPS will be allotted. The issue is later down in the Wiki entry (http://wiki.openvz.org/I/O_priorities_for_containers) is says this:
Quote: |
The mapping from priority to time is the following: if 0 prio corresponds to time slice t, than 8 prio corresponds to time slice 2 * t. Default time slice value is HZ/2. The main criteria for fairness at the moment is that if you set I/O prio of CT 1 to p1 and I/O prio of CT 2 to p2, and p1 > p2 then CT 1 should do more I/O than CT 2.
|
.. which seems to contradict the above statement. If VE1 has IO Priority of 1, and VE2 has IO Priority of 2 .. VE1 should be doing more IO than VE2.
Am I missing something or completely misunderstanding the Wiki entry?
Thanks!
[Updated on: Sun, 19 April 2009 18:47] Report message to a moderator
|
|
|