From: Herbert Poetzl (herbert_at_13thfloor.at)
Date: Thu 31 Oct 2002 - 02:03:08 GMT
On Wed, Oct 30, 2002 at 02:52:14PM -0700, Mefford, Aaron wrote:
> I realize that I am new to the list and do not have much history,
> but I am considering the possibility of using vserver on a fairly
> large project and as such would like to at least take a moment
> to offer my opinion on a couple of items.
I'm just curious: how did you find your way to the list?
> First, I almost walked away from using the vserver option until
> after joining the list I saw that the quota issue was being actively
> addressed. Of the todo's left, quota was the biggest gap for my
> application. It would be excellent to see the others addressed but
> without quota it would not be an option.
- so you actually require quota for your 'project'?
- would you like to help us with the quota issue?
- what about doing some testing?
> As to the specific post, I am not sure that the hard line of not
> overbooking is a good idea. While for many applications it
> would be a correct solution there are some where it will not.
> Every ISP over allocates their available resources.
unfortunately that's true ...
> People do not care to pay for dedicated resources.
hmm, I think that depends on the clientele ...
> Additionally, with most services now being offered via resellers,
> it seems unreasonable to not allow the reseller the same option.
> For instance, if I sell virtual private servers, and joe buys a
> VPS with the intention of selling individual web sites run within
> the VPS, I may or may not want to allow Joe to oversubscribe his
> disk space, possibly even on a per VPS basis.
> I realize that implementing a solution that would support a
> hybrid approach raises the complexity, but I wanted to state
> that there is value and need for such an approach.
what do you mean by hybrid approach?
- that you would be able to set quota or leave it unset?
- that you set the quota, but it might be ignored?
>>> - how to handle context quota violations within the kernel
>>> for users which do not exceed their personal quota?
>>> Simply report you exceeded your quota, and on check report
>>> that still space/quota is left?
>> IMHO, It should not be possible for a context to exceed it's quota when
>> some users have not. This is the point of quota mechanism. Guarantee
>> space on the disk and not allow for over-booking. Allocated user quota
>> should be subtracted from the total context quota, so that any users
>> with no quota should not be able to use that space. So in a way, users
hmm, that might be the original idea of quota, but
all current implementations do not guarantee, but only
limit the maximum available resources ...
if you want to guarantee, you then simply must do the
math an make sure that enough physical disk space is
available (or in the context case, the context quota
lies above the sum of all user quotas)
>> with quota will have their space, while other users will share what is
>> left. It is the only way to guarantee file allocation.
>> So, if context has 1Gig, and we allocate 300megs to users, all the other
>> users will get at max 700megs.