[OpenAFS-devel] Some tests and questions about fs setquota

[Derrick Brashear]

On Mon, Nov 24, 2008 at 9:51 AM, Jeffrey Altman
<jaltman@secure-endpoints.com> wrote:
> Derrick Brashear wrote:
>> On Mon, Nov 24, 2008 at 7:34 AM, Chas Williams (CONTRACTOR)
>> <chas@cmf.nrl.navy.mil> wrote:
>>> In message <4929C87A.9080200@rampaginggeek.com>,Jason Edgecombe writes:
>>>> If we want to allow a negative quota to disable all writes, then I
>>>> propose that we only allow -1. Then again, 1 is almost as effective
>>>> assuming that the volume has a few files in it.
>>> if you choose to go this route, then you will have a complete range of
>>> negative quotas and cut the quota range in half.  using a single bit from
>>> the value range is wasteful -- there is no reason for quota to be signed.
>
> Agreed.   The only problem is that the wire protocol defines the quota
> to be signed.  Changing to unsigned is a protocol change.
>
> When we add support for 64-bit volumes, then we should change the quota
> interface at the same time to support 64-bit quotas.
>
>> Agreed. A quota of 1 should be fine for "no write": the directory
>> object in an empty volume should be that big.
>
> Agreed.
>
>> I suppose I should try this.
>
> It will permit the ability to write 1MB.  We could special case the
> value 1 to mean do not permit the overage margin.

No, that'd be 1000 (or 1024 depending how you look at it). 1 is 1 *kb*


-- 
Derrick