Content-Type: text/plain;
	charset=us-ascii
Mime-Version: 1.0 (Mac OS X Mail 16.0 \(3864.700.51.1.3\))
From: Tim Bruijnzeels <tbruijnzeels@ripe.net>
In-Reply-To: 
 <178976427618.1287.16841443831092394455@dt-datatracker-fffdbdc97-wjd6q>
Date: Tue, 22 Sep 2026 13:25:54 +0200
Content-Transfer-Encoding: quoted-printable
Message-Id: <9F051FCC-D1B6-42FA-9F8A-63B87DAA3D7F@ripe.net>
References: 
 <178976427618.1287.16841443831092394455@dt-datatracker-fffdbdc97-wjd6q>
To: Roman Danyliw <rdd@cert.org>
Message-ID-Hash: CMB5QCRKFKAAP3GZJKT4JX5ZTZG564GU
CC: The IESG <iesg@ietf.org>,
 draft-ietf-sidrops-publication-server-bcp@ietf.org, ggx@gigix.net,
 sidrops-chairs@ietf.org, sidrops@ietf.org
Precedence: list
Subject: [Sidrops] Re: Roman Danyliw's Discuss on
 draft-ietf-sidrops-publication-server-bcp-11: (with DISCUSS and COMMENT)
Archived-At: 
 <https://mailarchive.ietf.org/arch/msg/sidrops/Dd-LRysbXggM0D3WXQD1bihJOzg>

Hi Roman, all,

> On 18 Sep 2026, at 22:44, Roman Danyliw via Datatracker =
<noreply@ietf.org> wrote:
>=20
> ** Section 6.3.
>   It is RECOMMENDED to serve rsync repositories from local storage, so
>   that the host operating system can optimally use its I/O cache.
>   Using network storage is NOT RECOMMENDED, because it may not benefit
>   from this cache.  For example, when using NFS, the operating system
>   might not be able to cache the directory listing(s) of the
>   repository.
>=20
> Is the WG confident that this generic statement considers all of the =
nuances of
> the very diverse set of network storage options in the marketplace.  =
Host OS
> I/O cache management is but one of many considerations.

To be honest, no...

This text was (of course) well intended and inspired by a real issue one
of the co-authors experienced. So, we believe it's good to raise =
awareness
but perhaps a tuned down statement is more appropriate.

How about:

   It is recommended to choose a storage backing that minimizes I/O
   overhead and allows the host operating system to fully leverage its
   filesystem cache. Operators should be cautious about using network
   storage, as doing so can have a negative impact on this and result
   in severely degraded rsync performance, especially on large
   repositories.

> If the WG confident it should be specifying hardware configurations =
for RPKI?

No, probably not.=20

Kind regards

Tim