[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CA+2bHPYYCj1rWyXqdPEVfbKhvueG9+BNXG-6-uQtzpPSD90jiQ@mail.gmail.com>
Date: Wed, 11 Sep 2024 14:03:49 -0400
From: Patrick Donnelly <pdonnell@...hat.com>
To: Max Kellermann <max.kellermann@...os.com>
Cc: xiubli@...hat.com, idryomov@...il.com, ceph-devel@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2] fs/ceph/quota: ignore quota with CAP_SYS_RESOURCE
On Wed, Sep 11, 2024 at 11:23 AM Max Kellermann
<max.kellermann@...os.com> wrote:
>
> On Wed, Sep 11, 2024 at 5:03 PM Patrick Donnelly <pdonnell@...hat.com> wrote:
> > Just because the client cooperatively maintains the quota limits with
> > the MDS does not mean it can override the quota in a distributed
> > system.
>
> I thought Ceph's quotas were implemented only on the client, just like
> file permissions. Is that not correct? Is there an additional
> server-side quota check? In my tests, I never saw one; it looked like
> I could write arbitrary amounts of data with my patch.
CephFS has many components that are cooperatively maintained by the
MDS **and** the clients; i.e. the clients are trusted to follow the
protocols and restrictions in the file system. For example,
capabilities grant a client read/write permissions on an inode but a
client could easily just open any file and write to it at will. There
is no barrier preventing that misbehavior.
Having root on a client does not extend to arbitrary superuser
permissions on the distributed file system. Down that path lies chaos
and inconsistency.
--
Patrick Donnelly, Ph.D.
He / Him / His
Red Hat Partner Engineer
IBM, Inc.
GPG: 19F28A586F808C2402351B93C3301A3E258DD79D
Powered by blists - more mailing lists