Hi Yufei,

Updating this based on the discussion in the Community Sync on Feb 19.

As discussed, [3395] complements the NoSQL Persistence implementation that
is already in `main` and allows Polaris Users to try out the NoSQL
persistence as a complete solution (even while it is flagged as "beta") in
the upcoming 1.4.0 release (main MetaStore, plus database maintenance tasks
required for controlling storage size).

Polaris as a project has historically been open to admitting experimental
features and allowing the to mature over a few subsequent releases.

The community is open to discussing / evolving the Admin Tool in general
after 1.4.0.

I believe you mentioned that end users could utilize nightly tool builds
for the NoSQL maintenance command. I'm not sure it is a viable solution, as
ASF policies [1] emphasise:

"Projects MUST direct outsiders towards official releases rather than raw
source repositories, nightly builds, snapshots, release candidates, or any
other similar packages."

All reviewers apparently support having [3395] in the Admin Tool at this
time. You seem to be the only person with open concerns. The community
values your opinion. Please give a recap (from your POV) of what might be a
disadvantage to merging [3395] today or tomorrow.

[1] https://www.apache.org/legal/release-policy.html#publication

[3395] https://github.com/apache/polaris/pull/3395

Thanks,
Dmitri.

On Tue, Feb 17, 2026 at 2:47 PM Dmitri Bourlatchkov <[email protected]>
wrote:

> Hi Yufei,
>
> I'm open to continuing the discussion on PR 3395 if needed, but I don't
> think it is a blocker for 1.4.0.
>
>
> It is not a blocker from the release perspective, but [3395] makes the
> NoSQL Persistence story "complete" from the end user's perspective.
>
> Given that everyone is in agreement with continuing discussions and
> improvements in the Admin Tool, would you mind merging [3395] specifically
> to provide a full solution to end users in 1.4.0 and adjusting on main with
> additional proposals / considerations later?
>
> I'm just trying to understand if there may be any disadvantage to merging
> it now. The possible "convenience" or "confusion" aspects can be discussed
> in docs and in any case are outweighed (from my POV) by the fact that users
> get a usable maintenance tool in the same release as the backend that needs
> this tool. My impression from previous discussions about this PR is that
> the majority of the community is fine with having it in the Admin Tool, at
> least as the next step in the evolution of the project.
>
> [3395] https://github.com/apache/polaris/pull/3395
>
> Thanks,
> Dmitri.
>
> On Tue, Feb 17, 2026 at 2:17 PM Yufei Gu <[email protected]> wrote:
>
>> +1 on what Adnan and JB proposed.
>>
>> Hi Dmitri,
>>
>> We need a few followups for PR 3409, like CLI and doc changes, but I think
>> it's fine to include it in 1.4.0 as it is guarded by the feature flag.
>>
>> I'm open to continuing the discussion on PR 3395 if needed, but I don't
>> think it is a blocker for 1.4.0.
>>
>> Yufei
>>
>>
>> On Tue, Feb 17, 2026 at 9:32 AM Dmitri Bourlatchkov <[email protected]>
>> wrote:
>>
>> > Hi All,
>> >
>> > Any concerns with merging [3409] before 1.4.0?
>> >
>> > Note: the new functionality is covered by a feature flag, which is off
>> by
>> > default.
>> >
>> > [3409] https://github.com/apache/polaris/pull/3409
>> >
>> > Thanks,
>> > Dmitri.
>> >
>> > On Mon, Feb 16, 2026 at 9:37 PM Adnan Hemani via dev <
>> > [email protected]>
>> > wrote:
>> >
>> > > Hi all,
>> > >
>> > > I've gone through the GH issues and PRs tagged to the 1.4.0 label and
>> > would
>> > > like to make the following recommendations for a potential Apache
>> Polaris
>> > > 1.4.0 release branch cut on (tentatively) 2026-02-23.
>> > >
>> > > Please reply to this thread prior to that date if there is any
>> feedback,
>> > > comments, and/or concerns so that we can get community consensus
>> before
>> > we
>> > > proceed with the release. If there are no replies to this email, the
>> > 1.4.0
>> > > release branch will be cut on 2026-02-23.
>> > >
>> > > Open Issues (Recommendation in []):
>> > > * [PUNT] #538 <https://github.com/apache/polaris/issues/538>: Table
>> > > Maintenance Support in Polaris
>> > >     * No major work in progress to justify delaying the release.
>> > >
>> > > * [CLOSE] #550 <https://github.com/apache/polaris/issues/550>:
>> Support
>> > for
>> > > GCP service account impersonation.
>> > >     * I will ping Michael to do this when he is back from vacation.
>> > >
>> > > * [CLOSE] #552 <https://github.com/apache/polaris/issues/552>: Safety
>> > > against unparseable locations.
>> > >     * I believe all the work has been completed for this. Dmitri, can
>> you
>> > > please confirm? (Will follow up offline as well)
>> > >
>> > > * [DISCUSS] #650 <https://github.com/apache/polaris/issues/650> /
>> #3395
>> > > <https://github.com/apache/polaris/pull/3395>: MongoDB Persistence
>> > Backend
>> > >     * It seems that discussions are still active and ongoing, and
>> there
>> > is
>> > > disagreement behind getting the change into Admin Tools while the core
>> > > functionality is merged already. I can see the arguments from both
>> sides
>> > to
>> > > push 1.4.0 without the Admin Tools change OR to hold the release until
>> > > there is agreement on these last bit of changes. What are the
>> community's
>> > > thoughts? Default option (if no one chimes in): we will push the
>> release
>> > > as-is on the 23rd.
>> > >
>> > > * [PUNT] #2671 <https://github.com/apache/polaris/issues/2671>: DB
>> > Schema
>> > > Migration Between Releases
>> > >     * No major work in progress to justify delaying the release.
>> > >
>> > > * [PUNT] #3685 <https://github.com/apache/polaris/issues/3685>:
>> > > `Create_Namespace` SQL Optimization
>> > >     * While there is traction, we may still be too far from a
>> load-tested
>> > > fix. This should be a high priority for 1.5.0.
>> > >
>> > > Open PRs:
>> > >
>> > > * [PUNT] #2180 <https://github.com/apache/polaris/pull/2180>: Async &
>> > > reliable tasks API, SPI, Store interfaces
>> > >     * PR is not very active over the last few weeks and does not have
>> > > enough active reviewers to see a strong path forward for merging in
>> the
>> > > next week or so.
>> > >
>> > > * [PUNT] #3256 <https://github.com/apache/polaris/pull/3256>: Object
>> > > Storage Operations
>> > >     * From the ML, it seems that there are still two rival proposals
>> that
>> > > are attempting to solve similar issues. My recommendation is to
>> unstick
>> > > this discussion from the 1.4.0 release to give proper time for the
>> > > discussion to resolve.
>> > >
>> > > When commenting, please reference the GH Issue/PR number so that we
>> are
>> > > clear on what is being discussed :)
>> > >
>> > > Best,
>> > > Adnan Hemani
>> > >
>> >
>>
>

Reply via email to