From: Filip Schauer <f.schauer@proxmox.com>
To: zedv@physik.fu-berlin.de,
Proxmox VE development discussion <pve-devel@lists.proxmox.com>
Subject: Re: [RFC container/storage] OCI image OverlayFS integration proposal
Date: Wed, 15 Jul 2026 12:00:27 +0200 [thread overview]
Message-ID: <571464e9-756b-4b80-9f1f-1b6905b247b5@proxmox.com> (raw)
In-Reply-To: <alZDjfWvqoQFKTV3@physik.fu-berlin.de>
On 14/07/2026 16:11, Jörg Behrmann wrote:
> Hi!
>
> I usually only watch from the sidelines, but this just tumbled through my mail
> client while I was waiting for a build to finish. I've not read the whole
> proposel, so excuse me if this comment is silly…
>
> On Tue, Jul 14, 2026 at 03:19:09PM +0200, Filip Schauer wrote:
>> Directory structure of the content type:
>> ```
>> oci/
>> +-- layers/
>> | +-- sha256-111.../
>> | +-- sha256-222.../
>> | `-- by-blob-id/
>> | +-- sha256-abc...
>> | `-- sha256-def...
>> `-- images/
>> +-- sha256-9f86d081...
>> `-- sha256-2c26b46b...
>> ```
>>
> This looks a bit like composefs. I grepped through your proposal and didn't see
> it mentioned.
>
> - The original C project https://github.com/composefs/composefs
> - The Rust rewrite https://github.com/composefs/composefs-rs
>
> RH built this because of their bootc plans. I'm not a fan of that and also not
> much a fan of OCI containers themselves, but I know the people working on it and
> they are nice and the project does have some cool uses.
>
> There's a talk from last year's All Systems Go at [1] and you can find the
> slides at [2]. The above reminded me particularly of slide 13 and later.
>
> I hope this isn't too far off from what you want to do.
>
> best regards,
> Jörg Behrmann
>
> [1] https://media.ccc.de/v/all-systems-go-2025-362-uki-composefs-and-remote-attestation-for-bootable-containers
> [2] https://cfp.all-systems-go.io/media/all-systems-go-2025/submissions/TNKPQS/resources/2025_All_Systems_Go_-_UKI_composefs_and_r9Z95aW.pdf
Hello,
thanks for pointing this out, composefs certainly looks interesting.
As I understand it, composefs deduplicates at the individual file level,
while my proposal deduplicates at the layer level.
composefs is not yet packaged in Debian stable. That said, it might be
something to keep in mind for a potential future extension.
For the initial version however, I would rather keep things simple and
stay aligned with the layer model defined by the OCI spec.
prev parent reply other threads:[~2026-07-15 10:00 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-07-14 13:19 [RFC container/storage] OCI image OverlayFS integration proposal Filip Schauer
[not found] ` <alZDjfWvqoQFKTV3@physik.fu-berlin.de>
2026-07-15 10:00 ` Filip Schauer [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=571464e9-756b-4b80-9f1f-1b6905b247b5@proxmox.com \
--to=f.schauer@proxmox.com \
--cc=pve-devel@lists.proxmox.com \
--cc=zedv@physik.fu-berlin.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.