l4-hurd
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Directories traversal (was Re: the deadly hypercube of death, or: ha


From: Marcus Brinkmann
Subject: Re: Directories traversal (was Re: the deadly hypercube of death, or: handling permissions)
Date: Fri, 28 Apr 2006 14:34:14 +0200
User-agent: Wanderlust/2.14.0 (Africa) SEMI/1.14.6 (Maruoka) FLIM/1.14.7 (Sanjō) APEL/10.6 Emacs/21.4 (i486-pc-linux-gnu) MULE/5.0 (SAKAKI)

At Fri, 28 Apr 2006 14:01:33 +0200,
Bas Wijnen <address@hidden> wrote:
> I think it's a very bad idea to give the system administrator read access to
> all your files, just because he wants to make backups.  There already is a
> form which can be used for recovering, because we have a persistent system.
> Making a backup should simply consist of copying the snapshot.  The question
> is who should have the right to do this, but it makes sense that there is at
> least a capability for it.
> 
> To get rid of the block device/files problem, only allocated parts of space
> banks should be backed up, and not the entire allocatable space.  Optimising
> more than that will lead to security and privacy issues, I think, and it's not
> important enough that we should risk that.

However, we will want a way to serialize state of many programs, and
transfer it to an updated set of servers, or maybe even to a new
machine (of course not possible if the state contains random
capabilities).

This mechanism could also be used to create backup of selected
configurations.  I think that this is important enough for partial
recovery that it must be considered.  Because you can not restore a
backup of the whole machine just because one user lost an important
file.

Of course, there are other solutions like versioned file systems.
This requires much more thought.

Thanks,
Marcus






reply via email to

[Prev in Thread] Current Thread [Next in Thread]