- From: Julian Reschke <julian.reschke@gmx.de>
- Date: Thu, 27 Mar 2003 20:49:22 +0100
- To: "WebDAV" <w3c-dist-auth@w3.org>
> From: w3c-dist-auth-request@w3.org > [mailto:w3c-dist-auth-request@w3.org]On Behalf Of Brian Korver > Sent: Thursday, March 27, 2003 1:05 AM > To: Julian Reschke > Cc: WebDAV > Subject: Re: I-D ACTION:draft-ietf-webdav-quota-01.txt > > > > On Tuesday, March 25, 2003, at 01:00 PM, Julian Reschke wrote: > > > > The reason I raised this is because I feel that a "disk full" indeed > > is an > > error condition on the server (thus 5xx), but a failed request due to > > exceeded quota limits happens *on purpose* -- the server theoretically > > *could* perform the request, but it doesn't want to. > > > > But as Geoff already stated, this is really not an important issue -- > > even > > more so if we have well-defined condition names that we can send in > > error > > response bodies. > > > > I'd prefer to focus the discussion on the other issues I mentioned > > (such as > > what is the quota model the spec describes, and do we really need to > > describe a specific model at all; or: are physical disk limits a > > special > > case of quotas? -- RFC3010 distinguishes both). > > > > Julian > > Julian, > > I was thinking that physical disk limits are just another class of "quotas". > That probably isn't clear enough from the spec, and should probably be > stated. Yes and no. It depends on who is asking. Note that RFC3010 (which I think we all like :-) distinguishes between both. Earlier today, Stefan E. showed that lumping disk limits and quota into the same property can lead to very surprising effects. Do we really want that? > The alternative is to report two numbers, quota-limit and something like > "space-limit", but I think we basically agreed that reporting back one number > would be better than multiple numbers (given that the server should know > which of the numbers is the "best" one to represent the space constraints > on that server). The issue here is that disk limits *behave* differently than quotas. For instance, a "quota error" usually can be resolved by the user, by deleting some of his own files, while a "disk full" condition may possibly not (other user may take up the disk space, so there's no way for the affected user to recover/continue). Also, recovery from both conditions may be different -- the person you need to contact to get a higher quota might be somebody else than the person that's able to free up physical disk space (or add new disks). Here's another proposal how we *could* inherit RFC3010 properties without much effort in WebDAV -- just use their welldefined names from RFC3010, and put them into the IETF namespace reserved for individual RFCs, for instance: <D:prop xmlns:D="DAV:" xmlns:NFS="urn:ietf:rfc:3010"> <NFS:space_freel>12345678</NFS:space_free> <NFS:space_total>1234567890</NFS:space_total> <NFS:quota_used>123456</NFS:quota_used> <NFS:quota_avail_hard>1234567</NFS:quota_avail_hard> </D:prop> (we'd just have to state how RFC3010 terminology maps to WebDAV resources) Julian -- <green/>bytes GmbH -- http://www.greenbytes.de -- tel:+492512807760
Received on Thursday, 27 March 2003 14:49:31 UTC