You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@madlib.apache.org by Dianjin Wang <wa...@gmail.com> on 2026/10/02 01:41:00 UTC
Re: [DISCUSS] Reviving Apache MADlib by re-establishing its PMC
Hi Niall,
Thank you for the quick and helpful reply, and for holding off on
the remaining Attic steps while we work out the best route.
We understand your point about the composition of the PMC. As a
next step, we will ask the Board whether a resolution to
re-establish MADlib is an option. We are also reaching out to
long-standing ASF members, including former MADlib PMC members,
to ask whether they would join the proposed PMC.
We will keep this thread updated on our progress.
Best,
Dianjin Wang
On Thu, Oct 1, 2026 at 5:40 AM Niall Pemberton
<ni...@gmail.com> wrote:
>
> Hi Dianjin
>
> On Wed, 30 Sep 2026 at 06:13, Dianjin Wang <dj...@apache.org> wrote:
>
> > Hi Attic team and all,
> >
> > As tison informed us, and as the GitBox listing now shows, Apache
> > MADlib has been retired and is being moved to the Attic [1]. We understand
> > the reasons: the PMC missed two roll calls and did not respond to
> > the final one. However, a small group of Apache Cloudberry
> > (Incubating) PPMC members who care about MADlib would like to take
> > on the responsibility of running the project again. We would like
> > to propose re-establishing the MADlib PMC.
> >
> > ## Why MADlib is worth reviving
> >
> > MADlib is still the only Apache project that provides in-database
> > machine learning and statistics for PostgreSQL-based engines. There
> > are no active committers, but people have kept contributing in the
> > past year:
> >
> > - Apache Cloudberry support (PR #627 [2])
> > - PostgreSQL 16/17/18 support
> > - CI migration from Jenkins to GitHub Actions
> > - Test Docker images updated to Ubuntu 22.04
> > - A new MkDocs-based website (madlib-site)
> >
> > All of this work is waiting for review and has not been merged.
> > The project is not stalled for lack of contributors. It is stalled
> > because no one with commit rights is left to review, merge and
> > release. We have described the direction in more detail on
> > dev@madlib [3][4].
> >
> > ## Proposed initial PMC
> >
> > - Leonid Borchuk (leborchuk@apache.org), Yandex Cloud
> > - Dianjin Wang (djwang@apache.org), HashData
> > - Max Yang (maxyang@apache.org), HashData
> >
> > All three of us serve on the Apache Cloudberry (Incubating) PPMC
> > and are familiar with how Apache projects are run. We know that
> > three members is the bare minimum and that two of us work for the
> > same company. Growing the PMC and making it more diverse is our top
> > priority. We will actively invite former MADlib committers,
> > PostgreSQL ecosystem users, and other contributors to join.
> >
> > ## What we commit to in the first 6-12 months
> >
> > 1. Review and merge the pending PRs, and publish an Apache release
> > (MADlib 2.2.0) with Cloudberry and PostgreSQL 16-18 support.
> > 2. Handle security reports and send board reports on time.
> > 3. Bring the build, CI and website back to a maintainable state.
> > 4. Grow the community by promoting MADlib in the PostgreSQL and
> > Cloudberry communities, lowering the barrier for new
> > contributors, and bringing in new committers and PMC members.
> >
> > If the community still fails to grow, our fallback is to propose
> > moving MADlib under the Apache Cloudberry PMC once Cloudberry
> > graduates. The project would stay at the ASF either way.
> >
> > ## Our requests
> >
> > 1. We would appreciate guidance on the process. Is a new board
> > resolution that re-establishes the PMC, naming the members above
> > and a VP, the right path? We are happy to draft it for an
> > upcoming board meeting.
>
>
> You would need to consult with the Board on whether they would accept a
> resolution to re-establish MADlib. I suspect they would only do this if the
> proposed PMC contained long standing ASF members. If the proposed PMC
> members have only been involved with an incubating project that hadn’t yet
> graduated, then the route to re-establishing MADlib would probably require
> re-entering the incubator. Alternatively, as you suggest, MADlib could
> become a sub-project of the Cloudberry incubating project.
>
> There is a paragraph on the Attic website (see “Process for Leaving the
> Attic” paragraph), but it doesn’t say much more than what I’ve said above:
>
> https://attic.apache.org/
>
>
> So the next steps could be to first ask the Board if a resolution to
> re-establish MADlib is an option. If they don’t accept that, then discuss
> with the incubator what the best option is (re-incubating MADlib as it’s on
> project or as a sub-project of Cloudberry).
>
> 2. While this is being discussed, could the project's assets
> > (repositories, website, JIRA, mailing list archives) be kept in
> > their current state, so that they can be restored easily if the
> > board approves?
>
>
> We haven’t yet started doing anything with MADlib. The last step is to ask
> Infra to make repos & JIRA read-only (which is reversible) and remove
> releases & mailing lists. You have some time before we start that & we can
> watch what happens while you work out the best route.
>
> Regards
>
> Niall
>
>
> > 3. We welcome any feedback or concerns from the Attic team, the
> > board, and former MADlib PMC members.
> >
> > Thank you for your time.
> >
> > [1] https://lists.apache.org/thread/zf4opv7omfcrdpr668j5bt9w8mf03m3l
> > [2] https://github.com/apache/madlib/pull/627
> > [3] https://lists.apache.org/thread/zdogpfdb6og54pnsohl99ojxkjc5c0zz
> > [4] https://lists.apache.org/thread/781rspl421vcmclzjdqlstzq8b5vgf4m
> >
> > Best,
> > Dianjin Wang
> >
Re: [DISCUSS] Reviving Apache MADlib by re-establishing its PMC
Posted by Dianjin Wang <wa...@gmail.com>.
Hi tison,
Thank you for discussing this with the directors and members, and
for finding a practical option that avoids a new resolution. We
appreciate the effort.
After thinking it over, we would still prefer to first try
re-establishing the MADlib PMC. Our reasons:
1. MADlib is used beyond Cloudberry. Many of its users and potential
contributors come from the wider PostgreSQL ecosystem. A fork
named cloudberry-madlib, maintained by the Cloudberry PPMC, may
suggest that MADlib now only serves Cloudberry. That could make
it harder to attract contributors from outside Cloudberry.
2. Re-establishing the PMC keeps the existing repositories, history,
open PRs, JIRA, website and mailing lists intact. Users and
contributors would keep the same entry points.
3. With you and Willem joining, the proposed PMC now includes two
ASF members and five people from three organizations. We hope
this addresses the main concern raised earlier.
So we would like to go ahead with the resolution to re-establish
the MADlib PMC, as discussed earlier, if you are still willing to
help prepare it. If the directors have specific concerns, we would
be glad to hear and address them.
If the Board decides that re-establishing the PMC is not
appropriate at this time, we will respect that decision. MADlib can
then complete its move to the Attic as planned. Separately, we plan
to start the graduation discussion for Apache Cloudberry soon. Once
Cloudberry graduates, we would propose, through the normal Attic
process, that the Cloudberry PMC take over MADlib as a sub-project.
We believe a graduated PMC would be a more stable home for MADlib
than a podling.
Best,
Dianjin Wang
On Sun, Oct 4, 2026 at 10:48 AM tison <wa...@gmail.com> wrote:
>
> After some discussion with directors and members, it seems we can have
> an easier path:
>
> 1. The Cloudberry PPMC can fork MADlib's code into cloudberry-madlib
> 2. The Cloudberry PPMC can release MADlib under the name MADlib
>
> Then we're done. No resolution is needed.
>
> We have a previous example where the Logging PMC took Flume's code and
> released org.apache.flume packages.
>
> The Attic should be fine to terminate MADlib and perhaps set up a link
> to Cloudberry's fork.
>
> Best,
> tison.
>
> Dianjin Wang <wa...@gmail.com> 于2026年10月3日周六 10:29写道：
> >
> > Hi Niall,
> >
> > I would like to share the latest progress.
> >
> > We received feedback from the board: ASF members need to be included
> > to form a new PMC. Luckily, now with the guidance and help of two ASF
> > members, Tison and Willem, we have a new PMC for MADlib:
> >
> > - Leonid Borchuk (leborchuk@apache.org), Yandex Cloud
> > - Dianjin Wang (djwang@apache.org), HashData
> > - Max Yang (maxyang@apache.org), HashData
> > - William Jiang (ningjiang@apache.org) (ASF member, Cloudberry mentor)
> > - Zili Chen (tison@apache.org) (ASF member)
> >
> > Tison will help prepare a resolution for the board meeting. So we
> > request that the Attic needs to help halt the MADlib move to the
> > Attic.
> >
> > If any questions, please let us know.
> >
> > Thanks!
> >
> > Best,
> > Dianjin Wang
> >
> > On Fri, Oct 2, 2026 at 9:41 AM Dianjin Wang <wa...@gmail.com> wrote:
> > >
> > > Hi Niall,
> > >
> > > Thank you for the quick and helpful reply, and for holding off on
> > > the remaining Attic steps while we work out the best route.
> > >
> > > We understand your point about the composition of the PMC. As a
> > > next step, we will ask the Board whether a resolution to
> > > re-establish MADlib is an option. We are also reaching out to
> > > long-standing ASF members, including former MADlib PMC members,
> > > to ask whether they would join the proposed PMC.
> > >
> > > We will keep this thread updated on our progress.
> > >
> > > Best,
> > > Dianjin Wang
> > >
> > > On Thu, Oct 1, 2026 at 5:40 AM Niall Pemberton
> > > <ni...@gmail.com> wrote:
> > > >
> > > > Hi Dianjin
> > > >
> > > > On Wed, 30 Sep 2026 at 06:13, Dianjin Wang <dj...@apache.org> wrote:
> > > >
> > > > > Hi Attic team and all,
> > > > >
> > > > > As tison informed us, and as the GitBox listing now shows, Apache
> > > > > MADlib has been retired and is being moved to the Attic [1]. We understand
> > > > > the reasons: the PMC missed two roll calls and did not respond to
> > > > > the final one. However, a small group of Apache Cloudberry
> > > > > (Incubating) PPMC members who care about MADlib would like to take
> > > > > on the responsibility of running the project again. We would like
> > > > > to propose re-establishing the MADlib PMC.
> > > > >
> > > > > ## Why MADlib is worth reviving
> > > > >
> > > > > MADlib is still the only Apache project that provides in-database
> > > > > machine learning and statistics for PostgreSQL-based engines. There
> > > > > are no active committers, but people have kept contributing in the
> > > > > past year:
> > > > >
> > > > > - Apache Cloudberry support (PR #627 [2])
> > > > > - PostgreSQL 16/17/18 support
> > > > > - CI migration from Jenkins to GitHub Actions
> > > > > - Test Docker images updated to Ubuntu 22.04
> > > > > - A new MkDocs-based website (madlib-site)
> > > > >
> > > > > All of this work is waiting for review and has not been merged.
> > > > > The project is not stalled for lack of contributors. It is stalled
> > > > > because no one with commit rights is left to review, merge and
> > > > > release. We have described the direction in more detail on
> > > > > dev@madlib [3][4].
> > > > >
> > > > > ## Proposed initial PMC
> > > > >
> > > > > - Leonid Borchuk (leborchuk@apache.org), Yandex Cloud
> > > > > - Dianjin Wang (djwang@apache.org), HashData
> > > > > - Max Yang (maxyang@apache.org), HashData
> > > > >
> > > > > All three of us serve on the Apache Cloudberry (Incubating) PPMC
> > > > > and are familiar with how Apache projects are run. We know that
> > > > > three members is the bare minimum and that two of us work for the
> > > > > same company. Growing the PMC and making it more diverse is our top
> > > > > priority. We will actively invite former MADlib committers,
> > > > > PostgreSQL ecosystem users, and other contributors to join.
> > > > >
> > > > > ## What we commit to in the first 6-12 months
> > > > >
> > > > > 1. Review and merge the pending PRs, and publish an Apache release
> > > > > (MADlib 2.2.0) with Cloudberry and PostgreSQL 16-18 support.
> > > > > 2. Handle security reports and send board reports on time.
> > > > > 3. Bring the build, CI and website back to a maintainable state.
> > > > > 4. Grow the community by promoting MADlib in the PostgreSQL and
> > > > > Cloudberry communities, lowering the barrier for new
> > > > > contributors, and bringing in new committers and PMC members.
> > > > >
> > > > > If the community still fails to grow, our fallback is to propose
> > > > > moving MADlib under the Apache Cloudberry PMC once Cloudberry
> > > > > graduates. The project would stay at the ASF either way.
> > > > >
> > > > > ## Our requests
> > > > >
> > > > > 1. We would appreciate guidance on the process. Is a new board
> > > > > resolution that re-establishes the PMC, naming the members above
> > > > > and a VP, the right path? We are happy to draft it for an
> > > > > upcoming board meeting.
> > > >
> > > >
> > > > You would need to consult with the Board on whether they would accept a
> > > > resolution to re-establish MADlib. I suspect they would only do this if the
> > > > proposed PMC contained long standing ASF members. If the proposed PMC
> > > > members have only been involved with an incubating project that hadn’t yet
> > > > graduated, then the route to re-establishing MADlib would probably require
> > > > re-entering the incubator. Alternatively, as you suggest, MADlib could
> > > > become a sub-project of the Cloudberry incubating project.
> > > >
> > > > There is a paragraph on the Attic website (see “Process for Leaving the
> > > > Attic” paragraph), but it doesn’t say much more than what I’ve said above:
> > > >
> > > > https://attic.apache.org/
> > > >
> > > >
> > > > So the next steps could be to first ask the Board if a resolution to
> > > > re-establish MADlib is an option. If they don’t accept that, then discuss
> > > > with the incubator what the best option is (re-incubating MADlib as it’s on
> > > > project or as a sub-project of Cloudberry).
> > > >
> > > > 2. While this is being discussed, could the project's assets
> > > > > (repositories, website, JIRA, mailing list archives) be kept in
> > > > > their current state, so that they can be restored easily if the
> > > > > board approves?
> > > >
> > > >
> > > > We haven’t yet started doing anything with MADlib. The last step is to ask
> > > > Infra to make repos & JIRA read-only (which is reversible) and remove
> > > > releases & mailing lists. You have some time before we start that & we can
> > > > watch what happens while you work out the best route.
> > > >
> > > > Regards
> > > >
> > > > Niall
> > > >
> > > >
> > > > > 3. We welcome any feedback or concerns from the Attic team, the
> > > > > board, and former MADlib PMC members.
> > > > >
> > > > > Thank you for your time.
> > > > >
> > > > > [1] https://lists.apache.org/thread/zf4opv7omfcrdpr668j5bt9w8mf03m3l
> > > > > [2] https://github.com/apache/madlib/pull/627
> > > > > [3] https://lists.apache.org/thread/zdogpfdb6og54pnsohl99ojxkjc5c0zz
> > > > > [4] https://lists.apache.org/thread/781rspl421vcmclzjdqlstzq8b5vgf4m
> > > > >
> > > > > Best,
> > > > > Dianjin Wang
> > > > >
Re: [DISCUSS] Reviving Apache MADlib by re-establishing its PMC
Posted by Dianjin Wang <wa...@gmail.com>.
Hi tison,
Thank you for discussing this with the directors and members, and
for finding a practical option that avoids a new resolution. We
appreciate the effort.
After thinking it over, we would still prefer to first try
re-establishing the MADlib PMC. Our reasons:
1. MADlib is used beyond Cloudberry. Many of its users and potential
contributors come from the wider PostgreSQL ecosystem. A fork
named cloudberry-madlib, maintained by the Cloudberry PPMC, may
suggest that MADlib now only serves Cloudberry. That could make
it harder to attract contributors from outside Cloudberry.
2. Re-establishing the PMC keeps the existing repositories, history,
open PRs, JIRA, website and mailing lists intact. Users and
contributors would keep the same entry points.
3. With you and Willem joining, the proposed PMC now includes two
ASF members and five people from three organizations. We hope
this addresses the main concern raised earlier.
So we would like to go ahead with the resolution to re-establish
the MADlib PMC, as discussed earlier, if you are still willing to
help prepare it. If the directors have specific concerns, we would
be glad to hear and address them.
If the Board decides that re-establishing the PMC is not
appropriate at this time, we will respect that decision. MADlib can
then complete its move to the Attic as planned. Separately, we plan
to start the graduation discussion for Apache Cloudberry soon. Once
Cloudberry graduates, we would propose, through the normal Attic
process, that the Cloudberry PMC take over MADlib as a sub-project.
We believe a graduated PMC would be a more stable home for MADlib
than a podling.
Best,
Dianjin Wang
On Sun, Oct 4, 2026 at 10:48 AM tison <wa...@gmail.com> wrote:
>
> After some discussion with directors and members, it seems we can have
> an easier path:
>
> 1. The Cloudberry PPMC can fork MADlib's code into cloudberry-madlib
> 2. The Cloudberry PPMC can release MADlib under the name MADlib
>
> Then we're done. No resolution is needed.
>
> We have a previous example where the Logging PMC took Flume's code and
> released org.apache.flume packages.
>
> The Attic should be fine to terminate MADlib and perhaps set up a link
> to Cloudberry's fork.
>
> Best,
> tison.
>
> Dianjin Wang <wa...@gmail.com> 于2026年10月3日周六 10:29写道：
> >
> > Hi Niall,
> >
> > I would like to share the latest progress.
> >
> > We received feedback from the board: ASF members need to be included
> > to form a new PMC. Luckily, now with the guidance and help of two ASF
> > members, Tison and Willem, we have a new PMC for MADlib:
> >
> > - Leonid Borchuk (leborchuk@apache.org), Yandex Cloud
> > - Dianjin Wang (djwang@apache.org), HashData
> > - Max Yang (maxyang@apache.org), HashData
> > - William Jiang (ningjiang@apache.org) (ASF member, Cloudberry mentor)
> > - Zili Chen (tison@apache.org) (ASF member)
> >
> > Tison will help prepare a resolution for the board meeting. So we
> > request that the Attic needs to help halt the MADlib move to the
> > Attic.
> >
> > If any questions, please let us know.
> >
> > Thanks!
> >
> > Best,
> > Dianjin Wang
> >
> > On Fri, Oct 2, 2026 at 9:41 AM Dianjin Wang <wa...@gmail.com> wrote:
> > >
> > > Hi Niall,
> > >
> > > Thank you for the quick and helpful reply, and for holding off on
> > > the remaining Attic steps while we work out the best route.
> > >
> > > We understand your point about the composition of the PMC. As a
> > > next step, we will ask the Board whether a resolution to
> > > re-establish MADlib is an option. We are also reaching out to
> > > long-standing ASF members, including former MADlib PMC members,
> > > to ask whether they would join the proposed PMC.
> > >
> > > We will keep this thread updated on our progress.
> > >
> > > Best,
> > > Dianjin Wang
> > >
> > > On Thu, Oct 1, 2026 at 5:40 AM Niall Pemberton
> > > <ni...@gmail.com> wrote:
> > > >
> > > > Hi Dianjin
> > > >
> > > > On Wed, 30 Sep 2026 at 06:13, Dianjin Wang <dj...@apache.org> wrote:
> > > >
> > > > > Hi Attic team and all,
> > > > >
> > > > > As tison informed us, and as the GitBox listing now shows, Apache
> > > > > MADlib has been retired and is being moved to the Attic [1]. We understand
> > > > > the reasons: the PMC missed two roll calls and did not respond to
> > > > > the final one. However, a small group of Apache Cloudberry
> > > > > (Incubating) PPMC members who care about MADlib would like to take
> > > > > on the responsibility of running the project again. We would like
> > > > > to propose re-establishing the MADlib PMC.
> > > > >
> > > > > ## Why MADlib is worth reviving
> > > > >
> > > > > MADlib is still the only Apache project that provides in-database
> > > > > machine learning and statistics for PostgreSQL-based engines. There
> > > > > are no active committers, but people have kept contributing in the
> > > > > past year:
> > > > >
> > > > > - Apache Cloudberry support (PR #627 [2])
> > > > > - PostgreSQL 16/17/18 support
> > > > > - CI migration from Jenkins to GitHub Actions
> > > > > - Test Docker images updated to Ubuntu 22.04
> > > > > - A new MkDocs-based website (madlib-site)
> > > > >
> > > > > All of this work is waiting for review and has not been merged.
> > > > > The project is not stalled for lack of contributors. It is stalled
> > > > > because no one with commit rights is left to review, merge and
> > > > > release. We have described the direction in more detail on
> > > > > dev@madlib [3][4].
> > > > >
> > > > > ## Proposed initial PMC
> > > > >
> > > > > - Leonid Borchuk (leborchuk@apache.org), Yandex Cloud
> > > > > - Dianjin Wang (djwang@apache.org), HashData
> > > > > - Max Yang (maxyang@apache.org), HashData
> > > > >
> > > > > All three of us serve on the Apache Cloudberry (Incubating) PPMC
> > > > > and are familiar with how Apache projects are run. We know that
> > > > > three members is the bare minimum and that two of us work for the
> > > > > same company. Growing the PMC and making it more diverse is our top
> > > > > priority. We will actively invite former MADlib committers,
> > > > > PostgreSQL ecosystem users, and other contributors to join.
> > > > >
> > > > > ## What we commit to in the first 6-12 months
> > > > >
> > > > > 1. Review and merge the pending PRs, and publish an Apache release
> > > > > (MADlib 2.2.0) with Cloudberry and PostgreSQL 16-18 support.
> > > > > 2. Handle security reports and send board reports on time.
> > > > > 3. Bring the build, CI and website back to a maintainable state.
> > > > > 4. Grow the community by promoting MADlib in the PostgreSQL and
> > > > > Cloudberry communities, lowering the barrier for new
> > > > > contributors, and bringing in new committers and PMC members.
> > > > >
> > > > > If the community still fails to grow, our fallback is to propose
> > > > > moving MADlib under the Apache Cloudberry PMC once Cloudberry
> > > > > graduates. The project would stay at the ASF either way.
> > > > >
> > > > > ## Our requests
> > > > >
> > > > > 1. We would appreciate guidance on the process. Is a new board
> > > > > resolution that re-establishes the PMC, naming the members above
> > > > > and a VP, the right path? We are happy to draft it for an
> > > > > upcoming board meeting.
> > > >
> > > >
> > > > You would need to consult with the Board on whether they would accept a
> > > > resolution to re-establish MADlib. I suspect they would only do this if the
> > > > proposed PMC contained long standing ASF members. If the proposed PMC
> > > > members have only been involved with an incubating project that hadn’t yet
> > > > graduated, then the route to re-establishing MADlib would probably require
> > > > re-entering the incubator. Alternatively, as you suggest, MADlib could
> > > > become a sub-project of the Cloudberry incubating project.
> > > >
> > > > There is a paragraph on the Attic website (see “Process for Leaving the
> > > > Attic” paragraph), but it doesn’t say much more than what I’ve said above:
> > > >
> > > > https://attic.apache.org/
> > > >
> > > >
> > > > So the next steps could be to first ask the Board if a resolution to
> > > > re-establish MADlib is an option. If they don’t accept that, then discuss
> > > > with the incubator what the best option is (re-incubating MADlib as it’s on
> > > > project or as a sub-project of Cloudberry).
> > > >
> > > > 2. While this is being discussed, could the project's assets
> > > > > (repositories, website, JIRA, mailing list archives) be kept in
> > > > > their current state, so that they can be restored easily if the
> > > > > board approves?
> > > >
> > > >
> > > > We haven’t yet started doing anything with MADlib. The last step is to ask
> > > > Infra to make repos & JIRA read-only (which is reversible) and remove
> > > > releases & mailing lists. You have some time before we start that & we can
> > > > watch what happens while you work out the best route.
> > > >
> > > > Regards
> > > >
> > > > Niall
> > > >
> > > >
> > > > > 3. We welcome any feedback or concerns from the Attic team, the
> > > > > board, and former MADlib PMC members.
> > > > >
> > > > > Thank you for your time.
> > > > >
> > > > > [1] https://lists.apache.org/thread/zf4opv7omfcrdpr668j5bt9w8mf03m3l
> > > > > [2] https://github.com/apache/madlib/pull/627
> > > > > [3] https://lists.apache.org/thread/zdogpfdb6og54pnsohl99ojxkjc5c0zz
> > > > > [4] https://lists.apache.org/thread/781rspl421vcmclzjdqlstzq8b5vgf4m
> > > > >
> > > > > Best,
> > > > > Dianjin Wang
> > > > >
Re: [DISCUSS] Reviving Apache MADlib by re-establishing its PMC
Posted by tison <wa...@gmail.com>.
After some discussion with directors and members, it seems we can have
an easier path:
1. The Cloudberry PPMC can fork MADlib's code into cloudberry-madlib
2. The Cloudberry PPMC can release MADlib under the name MADlib
Then we're done. No resolution is needed.
We have a previous example where the Logging PMC took Flume's code and
released org.apache.flume packages.
The Attic should be fine to terminate MADlib and perhaps set up a link
to Cloudberry's fork.
Best,
tison.
Dianjin Wang <wa...@gmail.com> 于2026年10月3日周六 10:29写道：
>
> Hi Niall,
>
> I would like to share the latest progress.
>
> We received feedback from the board: ASF members need to be included
> to form a new PMC. Luckily, now with the guidance and help of two ASF
> members, Tison and Willem, we have a new PMC for MADlib:
>
> - Leonid Borchuk (leborchuk@apache.org), Yandex Cloud
> - Dianjin Wang (djwang@apache.org), HashData
> - Max Yang (maxyang@apache.org), HashData
> - William Jiang (ningjiang@apache.org) (ASF member, Cloudberry mentor)
> - Zili Chen (tison@apache.org) (ASF member)
>
> Tison will help prepare a resolution for the board meeting. So we
> request that the Attic needs to help halt the MADlib move to the
> Attic.
>
> If any questions, please let us know.
>
> Thanks!
>
> Best,
> Dianjin Wang
>
> On Fri, Oct 2, 2026 at 9:41 AM Dianjin Wang <wa...@gmail.com> wrote:
> >
> > Hi Niall,
> >
> > Thank you for the quick and helpful reply, and for holding off on
> > the remaining Attic steps while we work out the best route.
> >
> > We understand your point about the composition of the PMC. As a
> > next step, we will ask the Board whether a resolution to
> > re-establish MADlib is an option. We are also reaching out to
> > long-standing ASF members, including former MADlib PMC members,
> > to ask whether they would join the proposed PMC.
> >
> > We will keep this thread updated on our progress.
> >
> > Best,
> > Dianjin Wang
> >
> > On Thu, Oct 1, 2026 at 5:40 AM Niall Pemberton
> > <ni...@gmail.com> wrote:
> > >
> > > Hi Dianjin
> > >
> > > On Wed, 30 Sep 2026 at 06:13, Dianjin Wang <dj...@apache.org> wrote:
> > >
> > > > Hi Attic team and all,
> > > >
> > > > As tison informed us, and as the GitBox listing now shows, Apache
> > > > MADlib has been retired and is being moved to the Attic [1]. We understand
> > > > the reasons: the PMC missed two roll calls and did not respond to
> > > > the final one. However, a small group of Apache Cloudberry
> > > > (Incubating) PPMC members who care about MADlib would like to take
> > > > on the responsibility of running the project again. We would like
> > > > to propose re-establishing the MADlib PMC.
> > > >
> > > > ## Why MADlib is worth reviving
> > > >
> > > > MADlib is still the only Apache project that provides in-database
> > > > machine learning and statistics for PostgreSQL-based engines. There
> > > > are no active committers, but people have kept contributing in the
> > > > past year:
> > > >
> > > > - Apache Cloudberry support (PR #627 [2])
> > > > - PostgreSQL 16/17/18 support
> > > > - CI migration from Jenkins to GitHub Actions
> > > > - Test Docker images updated to Ubuntu 22.04
> > > > - A new MkDocs-based website (madlib-site)
> > > >
> > > > All of this work is waiting for review and has not been merged.
> > > > The project is not stalled for lack of contributors. It is stalled
> > > > because no one with commit rights is left to review, merge and
> > > > release. We have described the direction in more detail on
> > > > dev@madlib [3][4].
> > > >
> > > > ## Proposed initial PMC
> > > >
> > > > - Leonid Borchuk (leborchuk@apache.org), Yandex Cloud
> > > > - Dianjin Wang (djwang@apache.org), HashData
> > > > - Max Yang (maxyang@apache.org), HashData
> > > >
> > > > All three of us serve on the Apache Cloudberry (Incubating) PPMC
> > > > and are familiar with how Apache projects are run. We know that
> > > > three members is the bare minimum and that two of us work for the
> > > > same company. Growing the PMC and making it more diverse is our top
> > > > priority. We will actively invite former MADlib committers,
> > > > PostgreSQL ecosystem users, and other contributors to join.
> > > >
> > > > ## What we commit to in the first 6-12 months
> > > >
> > > > 1. Review and merge the pending PRs, and publish an Apache release
> > > > (MADlib 2.2.0) with Cloudberry and PostgreSQL 16-18 support.
> > > > 2. Handle security reports and send board reports on time.
> > > > 3. Bring the build, CI and website back to a maintainable state.
> > > > 4. Grow the community by promoting MADlib in the PostgreSQL and
> > > > Cloudberry communities, lowering the barrier for new
> > > > contributors, and bringing in new committers and PMC members.
> > > >
> > > > If the community still fails to grow, our fallback is to propose
> > > > moving MADlib under the Apache Cloudberry PMC once Cloudberry
> > > > graduates. The project would stay at the ASF either way.
> > > >
> > > > ## Our requests
> > > >
> > > > 1. We would appreciate guidance on the process. Is a new board
> > > > resolution that re-establishes the PMC, naming the members above
> > > > and a VP, the right path? We are happy to draft it for an
> > > > upcoming board meeting.
> > >
> > >
> > > You would need to consult with the Board on whether they would accept a
> > > resolution to re-establish MADlib. I suspect they would only do this if the
> > > proposed PMC contained long standing ASF members. If the proposed PMC
> > > members have only been involved with an incubating project that hadn’t yet
> > > graduated, then the route to re-establishing MADlib would probably require
> > > re-entering the incubator. Alternatively, as you suggest, MADlib could
> > > become a sub-project of the Cloudberry incubating project.
> > >
> > > There is a paragraph on the Attic website (see “Process for Leaving the
> > > Attic” paragraph), but it doesn’t say much more than what I’ve said above:
> > >
> > > https://attic.apache.org/
> > >
> > >
> > > So the next steps could be to first ask the Board if a resolution to
> > > re-establish MADlib is an option. If they don’t accept that, then discuss
> > > with the incubator what the best option is (re-incubating MADlib as it’s on
> > > project or as a sub-project of Cloudberry).
> > >
> > > 2. While this is being discussed, could the project's assets
> > > > (repositories, website, JIRA, mailing list archives) be kept in
> > > > their current state, so that they can be restored easily if the
> > > > board approves?
> > >
> > >
> > > We haven’t yet started doing anything with MADlib. The last step is to ask
> > > Infra to make repos & JIRA read-only (which is reversible) and remove
> > > releases & mailing lists. You have some time before we start that & we can
> > > watch what happens while you work out the best route.
> > >
> > > Regards
> > >
> > > Niall
> > >
> > >
> > > > 3. We welcome any feedback or concerns from the Attic team, the
> > > > board, and former MADlib PMC members.
> > > >
> > > > Thank you for your time.
> > > >
> > > > [1] https://lists.apache.org/thread/zf4opv7omfcrdpr668j5bt9w8mf03m3l
> > > > [2] https://github.com/apache/madlib/pull/627
> > > > [3] https://lists.apache.org/thread/zdogpfdb6og54pnsohl99ojxkjc5c0zz
> > > > [4] https://lists.apache.org/thread/781rspl421vcmclzjdqlstzq8b5vgf4m
> > > >
> > > > Best,
> > > > Dianjin Wang
> > > >
Re: [DISCUSS] Reviving Apache MADlib by re-establishing its PMC
Posted by tison <wa...@gmail.com>.
After some discussion with directors and members, it seems we can have
an easier path:
1. The Cloudberry PPMC can fork MADlib's code into cloudberry-madlib
2. The Cloudberry PPMC can release MADlib under the name MADlib
Then we're done. No resolution is needed.
We have a previous example where the Logging PMC took Flume's code and
released org.apache.flume packages.
The Attic should be fine to terminate MADlib and perhaps set up a link
to Cloudberry's fork.
Best,
tison.
Dianjin Wang <wa...@gmail.com> 于2026年10月3日周六 10:29写道：
>
> Hi Niall,
>
> I would like to share the latest progress.
>
> We received feedback from the board: ASF members need to be included
> to form a new PMC. Luckily, now with the guidance and help of two ASF
> members, Tison and Willem, we have a new PMC for MADlib:
>
> - Leonid Borchuk (leborchuk@apache.org), Yandex Cloud
> - Dianjin Wang (djwang@apache.org), HashData
> - Max Yang (maxyang@apache.org), HashData
> - William Jiang (ningjiang@apache.org) (ASF member, Cloudberry mentor)
> - Zili Chen (tison@apache.org) (ASF member)
>
> Tison will help prepare a resolution for the board meeting. So we
> request that the Attic needs to help halt the MADlib move to the
> Attic.
>
> If any questions, please let us know.
>
> Thanks!
>
> Best,
> Dianjin Wang
>
> On Fri, Oct 2, 2026 at 9:41 AM Dianjin Wang <wa...@gmail.com> wrote:
> >
> > Hi Niall,
> >
> > Thank you for the quick and helpful reply, and for holding off on
> > the remaining Attic steps while we work out the best route.
> >
> > We understand your point about the composition of the PMC. As a
> > next step, we will ask the Board whether a resolution to
> > re-establish MADlib is an option. We are also reaching out to
> > long-standing ASF members, including former MADlib PMC members,
> > to ask whether they would join the proposed PMC.
> >
> > We will keep this thread updated on our progress.
> >
> > Best,
> > Dianjin Wang
> >
> > On Thu, Oct 1, 2026 at 5:40 AM Niall Pemberton
> > <ni...@gmail.com> wrote:
> > >
> > > Hi Dianjin
> > >
> > > On Wed, 30 Sep 2026 at 06:13, Dianjin Wang <dj...@apache.org> wrote:
> > >
> > > > Hi Attic team and all,
> > > >
> > > > As tison informed us, and as the GitBox listing now shows, Apache
> > > > MADlib has been retired and is being moved to the Attic [1]. We understand
> > > > the reasons: the PMC missed two roll calls and did not respond to
> > > > the final one. However, a small group of Apache Cloudberry
> > > > (Incubating) PPMC members who care about MADlib would like to take
> > > > on the responsibility of running the project again. We would like
> > > > to propose re-establishing the MADlib PMC.
> > > >
> > > > ## Why MADlib is worth reviving
> > > >
> > > > MADlib is still the only Apache project that provides in-database
> > > > machine learning and statistics for PostgreSQL-based engines. There
> > > > are no active committers, but people have kept contributing in the
> > > > past year:
> > > >
> > > > - Apache Cloudberry support (PR #627 [2])
> > > > - PostgreSQL 16/17/18 support
> > > > - CI migration from Jenkins to GitHub Actions
> > > > - Test Docker images updated to Ubuntu 22.04
> > > > - A new MkDocs-based website (madlib-site)
> > > >
> > > > All of this work is waiting for review and has not been merged.
> > > > The project is not stalled for lack of contributors. It is stalled
> > > > because no one with commit rights is left to review, merge and
> > > > release. We have described the direction in more detail on
> > > > dev@madlib [3][4].
> > > >
> > > > ## Proposed initial PMC
> > > >
> > > > - Leonid Borchuk (leborchuk@apache.org), Yandex Cloud
> > > > - Dianjin Wang (djwang@apache.org), HashData
> > > > - Max Yang (maxyang@apache.org), HashData
> > > >
> > > > All three of us serve on the Apache Cloudberry (Incubating) PPMC
> > > > and are familiar with how Apache projects are run. We know that
> > > > three members is the bare minimum and that two of us work for the
> > > > same company. Growing the PMC and making it more diverse is our top
> > > > priority. We will actively invite former MADlib committers,
> > > > PostgreSQL ecosystem users, and other contributors to join.
> > > >
> > > > ## What we commit to in the first 6-12 months
> > > >
> > > > 1. Review and merge the pending PRs, and publish an Apache release
> > > > (MADlib 2.2.0) with Cloudberry and PostgreSQL 16-18 support.
> > > > 2. Handle security reports and send board reports on time.
> > > > 3. Bring the build, CI and website back to a maintainable state.
> > > > 4. Grow the community by promoting MADlib in the PostgreSQL and
> > > > Cloudberry communities, lowering the barrier for new
> > > > contributors, and bringing in new committers and PMC members.
> > > >
> > > > If the community still fails to grow, our fallback is to propose
> > > > moving MADlib under the Apache Cloudberry PMC once Cloudberry
> > > > graduates. The project would stay at the ASF either way.
> > > >
> > > > ## Our requests
> > > >
> > > > 1. We would appreciate guidance on the process. Is a new board
> > > > resolution that re-establishes the PMC, naming the members above
> > > > and a VP, the right path? We are happy to draft it for an
> > > > upcoming board meeting.
> > >
> > >
> > > You would need to consult with the Board on whether they would accept a
> > > resolution to re-establish MADlib. I suspect they would only do this if the
> > > proposed PMC contained long standing ASF members. If the proposed PMC
> > > members have only been involved with an incubating project that hadn’t yet
> > > graduated, then the route to re-establishing MADlib would probably require
> > > re-entering the incubator. Alternatively, as you suggest, MADlib could
> > > become a sub-project of the Cloudberry incubating project.
> > >
> > > There is a paragraph on the Attic website (see “Process for Leaving the
> > > Attic” paragraph), but it doesn’t say much more than what I’ve said above:
> > >
> > > https://attic.apache.org/
> > >
> > >
> > > So the next steps could be to first ask the Board if a resolution to
> > > re-establish MADlib is an option. If they don’t accept that, then discuss
> > > with the incubator what the best option is (re-incubating MADlib as it’s on
> > > project or as a sub-project of Cloudberry).
> > >
> > > 2. While this is being discussed, could the project's assets
> > > > (repositories, website, JIRA, mailing list archives) be kept in
> > > > their current state, so that they can be restored easily if the
> > > > board approves?
> > >
> > >
> > > We haven’t yet started doing anything with MADlib. The last step is to ask
> > > Infra to make repos & JIRA read-only (which is reversible) and remove
> > > releases & mailing lists. You have some time before we start that & we can
> > > watch what happens while you work out the best route.
> > >
> > > Regards
> > >
> > > Niall
> > >
> > >
> > > > 3. We welcome any feedback or concerns from the Attic team, the
> > > > board, and former MADlib PMC members.
> > > >
> > > > Thank you for your time.
> > > >
> > > > [1] https://lists.apache.org/thread/zf4opv7omfcrdpr668j5bt9w8mf03m3l
> > > > [2] https://github.com/apache/madlib/pull/627
> > > > [3] https://lists.apache.org/thread/zdogpfdb6og54pnsohl99ojxkjc5c0zz
> > > > [4] https://lists.apache.org/thread/781rspl421vcmclzjdqlstzq8b5vgf4m
> > > >
> > > > Best,
> > > > Dianjin Wang
> > > >
Re: [DISCUSS] Reviving Apache MADlib by re-establishing its PMC
Posted by Dianjin Wang <wa...@gmail.com>.
Hi Niall,
I would like to share the latest progress.
We received feedback from the board: ASF members need to be included
to form a new PMC. Luckily, now with the guidance and help of two ASF
members, Tison and Willem, we have a new PMC for MADlib:
- Leonid Borchuk (leborchuk@apache.org), Yandex Cloud
- Dianjin Wang (djwang@apache.org), HashData
- Max Yang (maxyang@apache.org), HashData
- William Jiang (ningjiang@apache.org) (ASF member, Cloudberry mentor)
- Zili Chen (tison@apache.org) (ASF member)
Tison will help prepare a resolution for the board meeting. So we
request that the Attic needs to help halt the MADlib move to the
Attic.
If any questions, please let us know.
Thanks!
Best,
Dianjin Wang
On Fri, Oct 2, 2026 at 9:41 AM Dianjin Wang <wa...@gmail.com> wrote:
>
> Hi Niall,
>
> Thank you for the quick and helpful reply, and for holding off on
> the remaining Attic steps while we work out the best route.
>
> We understand your point about the composition of the PMC. As a
> next step, we will ask the Board whether a resolution to
> re-establish MADlib is an option. We are also reaching out to
> long-standing ASF members, including former MADlib PMC members,
> to ask whether they would join the proposed PMC.
>
> We will keep this thread updated on our progress.
>
> Best,
> Dianjin Wang
>
> On Thu, Oct 1, 2026 at 5:40 AM Niall Pemberton
> <ni...@gmail.com> wrote:
> >
> > Hi Dianjin
> >
> > On Wed, 30 Sep 2026 at 06:13, Dianjin Wang <dj...@apache.org> wrote:
> >
> > > Hi Attic team and all,
> > >
> > > As tison informed us, and as the GitBox listing now shows, Apache
> > > MADlib has been retired and is being moved to the Attic [1]. We understand
> > > the reasons: the PMC missed two roll calls and did not respond to
> > > the final one. However, a small group of Apache Cloudberry
> > > (Incubating) PPMC members who care about MADlib would like to take
> > > on the responsibility of running the project again. We would like
> > > to propose re-establishing the MADlib PMC.
> > >
> > > ## Why MADlib is worth reviving
> > >
> > > MADlib is still the only Apache project that provides in-database
> > > machine learning and statistics for PostgreSQL-based engines. There
> > > are no active committers, but people have kept contributing in the
> > > past year:
> > >
> > > - Apache Cloudberry support (PR #627 [2])
> > > - PostgreSQL 16/17/18 support
> > > - CI migration from Jenkins to GitHub Actions
> > > - Test Docker images updated to Ubuntu 22.04
> > > - A new MkDocs-based website (madlib-site)
> > >
> > > All of this work is waiting for review and has not been merged.
> > > The project is not stalled for lack of contributors. It is stalled
> > > because no one with commit rights is left to review, merge and
> > > release. We have described the direction in more detail on
> > > dev@madlib [3][4].
> > >
> > > ## Proposed initial PMC
> > >
> > > - Leonid Borchuk (leborchuk@apache.org), Yandex Cloud
> > > - Dianjin Wang (djwang@apache.org), HashData
> > > - Max Yang (maxyang@apache.org), HashData
> > >
> > > All three of us serve on the Apache Cloudberry (Incubating) PPMC
> > > and are familiar with how Apache projects are run. We know that
> > > three members is the bare minimum and that two of us work for the
> > > same company. Growing the PMC and making it more diverse is our top
> > > priority. We will actively invite former MADlib committers,
> > > PostgreSQL ecosystem users, and other contributors to join.
> > >
> > > ## What we commit to in the first 6-12 months
> > >
> > > 1. Review and merge the pending PRs, and publish an Apache release
> > > (MADlib 2.2.0) with Cloudberry and PostgreSQL 16-18 support.
> > > 2. Handle security reports and send board reports on time.
> > > 3. Bring the build, CI and website back to a maintainable state.
> > > 4. Grow the community by promoting MADlib in the PostgreSQL and
> > > Cloudberry communities, lowering the barrier for new
> > > contributors, and bringing in new committers and PMC members.
> > >
> > > If the community still fails to grow, our fallback is to propose
> > > moving MADlib under the Apache Cloudberry PMC once Cloudberry
> > > graduates. The project would stay at the ASF either way.
> > >
> > > ## Our requests
> > >
> > > 1. We would appreciate guidance on the process. Is a new board
> > > resolution that re-establishes the PMC, naming the members above
> > > and a VP, the right path? We are happy to draft it for an
> > > upcoming board meeting.
> >
> >
> > You would need to consult with the Board on whether they would accept a
> > resolution to re-establish MADlib. I suspect they would only do this if the
> > proposed PMC contained long standing ASF members. If the proposed PMC
> > members have only been involved with an incubating project that hadn’t yet
> > graduated, then the route to re-establishing MADlib would probably require
> > re-entering the incubator. Alternatively, as you suggest, MADlib could
> > become a sub-project of the Cloudberry incubating project.
> >
> > There is a paragraph on the Attic website (see “Process for Leaving the
> > Attic” paragraph), but it doesn’t say much more than what I’ve said above:
> >
> > https://attic.apache.org/
> >
> >
> > So the next steps could be to first ask the Board if a resolution to
> > re-establish MADlib is an option. If they don’t accept that, then discuss
> > with the incubator what the best option is (re-incubating MADlib as it’s on
> > project or as a sub-project of Cloudberry).
> >
> > 2. While this is being discussed, could the project's assets
> > > (repositories, website, JIRA, mailing list archives) be kept in
> > > their current state, so that they can be restored easily if the
> > > board approves?
> >
> >
> > We haven’t yet started doing anything with MADlib. The last step is to ask
> > Infra to make repos & JIRA read-only (which is reversible) and remove
> > releases & mailing lists. You have some time before we start that & we can
> > watch what happens while you work out the best route.
> >
> > Regards
> >
> > Niall
> >
> >
> > > 3. We welcome any feedback or concerns from the Attic team, the
> > > board, and former MADlib PMC members.
> > >
> > > Thank you for your time.
> > >
> > > [1] https://lists.apache.org/thread/zf4opv7omfcrdpr668j5bt9w8mf03m3l
> > > [2] https://github.com/apache/madlib/pull/627
> > > [3] https://lists.apache.org/thread/zdogpfdb6og54pnsohl99ojxkjc5c0zz
> > > [4] https://lists.apache.org/thread/781rspl421vcmclzjdqlstzq8b5vgf4m
> > >
> > > Best,
> > > Dianjin Wang
> > >
Re: [DISCUSS] Reviving Apache MADlib by re-establishing its PMC
Posted by Dianjin Wang <wa...@gmail.com>.
Hi Niall,
I would like to share the latest progress.
We received feedback from the board: ASF members need to be included
to form a new PMC. Luckily, now with the guidance and help of two ASF
members, Tison and Willem, we have a new PMC for MADlib:
- Leonid Borchuk (leborchuk@apache.org), Yandex Cloud
- Dianjin Wang (djwang@apache.org), HashData
- Max Yang (maxyang@apache.org), HashData
- William Jiang (ningjiang@apache.org) (ASF member, Cloudberry mentor)
- Zili Chen (tison@apache.org) (ASF member)
Tison will help prepare a resolution for the board meeting. So we
request that the Attic needs to help halt the MADlib move to the
Attic.
If any questions, please let us know.
Thanks!
Best,
Dianjin Wang
On Fri, Oct 2, 2026 at 9:41 AM Dianjin Wang <wa...@gmail.com> wrote:
>
> Hi Niall,
>
> Thank you for the quick and helpful reply, and for holding off on
> the remaining Attic steps while we work out the best route.
>
> We understand your point about the composition of the PMC. As a
> next step, we will ask the Board whether a resolution to
> re-establish MADlib is an option. We are also reaching out to
> long-standing ASF members, including former MADlib PMC members,
> to ask whether they would join the proposed PMC.
>
> We will keep this thread updated on our progress.
>
> Best,
> Dianjin Wang
>
> On Thu, Oct 1, 2026 at 5:40 AM Niall Pemberton
> <ni...@gmail.com> wrote:
> >
> > Hi Dianjin
> >
> > On Wed, 30 Sep 2026 at 06:13, Dianjin Wang <dj...@apache.org> wrote:
> >
> > > Hi Attic team and all,
> > >
> > > As tison informed us, and as the GitBox listing now shows, Apache
> > > MADlib has been retired and is being moved to the Attic [1]. We understand
> > > the reasons: the PMC missed two roll calls and did not respond to
> > > the final one. However, a small group of Apache Cloudberry
> > > (Incubating) PPMC members who care about MADlib would like to take
> > > on the responsibility of running the project again. We would like
> > > to propose re-establishing the MADlib PMC.
> > >
> > > ## Why MADlib is worth reviving
> > >
> > > MADlib is still the only Apache project that provides in-database
> > > machine learning and statistics for PostgreSQL-based engines. There
> > > are no active committers, but people have kept contributing in the
> > > past year:
> > >
> > > - Apache Cloudberry support (PR #627 [2])
> > > - PostgreSQL 16/17/18 support
> > > - CI migration from Jenkins to GitHub Actions
> > > - Test Docker images updated to Ubuntu 22.04
> > > - A new MkDocs-based website (madlib-site)
> > >
> > > All of this work is waiting for review and has not been merged.
> > > The project is not stalled for lack of contributors. It is stalled
> > > because no one with commit rights is left to review, merge and
> > > release. We have described the direction in more detail on
> > > dev@madlib [3][4].
> > >
> > > ## Proposed initial PMC
> > >
> > > - Leonid Borchuk (leborchuk@apache.org), Yandex Cloud
> > > - Dianjin Wang (djwang@apache.org), HashData
> > > - Max Yang (maxyang@apache.org), HashData
> > >
> > > All three of us serve on the Apache Cloudberry (Incubating) PPMC
> > > and are familiar with how Apache projects are run. We know that
> > > three members is the bare minimum and that two of us work for the
> > > same company. Growing the PMC and making it more diverse is our top
> > > priority. We will actively invite former MADlib committers,
> > > PostgreSQL ecosystem users, and other contributors to join.
> > >
> > > ## What we commit to in the first 6-12 months
> > >
> > > 1. Review and merge the pending PRs, and publish an Apache release
> > > (MADlib 2.2.0) with Cloudberry and PostgreSQL 16-18 support.
> > > 2. Handle security reports and send board reports on time.
> > > 3. Bring the build, CI and website back to a maintainable state.
> > > 4. Grow the community by promoting MADlib in the PostgreSQL and
> > > Cloudberry communities, lowering the barrier for new
> > > contributors, and bringing in new committers and PMC members.
> > >
> > > If the community still fails to grow, our fallback is to propose
> > > moving MADlib under the Apache Cloudberry PMC once Cloudberry
> > > graduates. The project would stay at the ASF either way.
> > >
> > > ## Our requests
> > >
> > > 1. We would appreciate guidance on the process. Is a new board
> > > resolution that re-establishes the PMC, naming the members above
> > > and a VP, the right path? We are happy to draft it for an
> > > upcoming board meeting.
> >
> >
> > You would need to consult with the Board on whether they would accept a
> > resolution to re-establish MADlib. I suspect they would only do this if the
> > proposed PMC contained long standing ASF members. If the proposed PMC
> > members have only been involved with an incubating project that hadn’t yet
> > graduated, then the route to re-establishing MADlib would probably require
> > re-entering the incubator. Alternatively, as you suggest, MADlib could
> > become a sub-project of the Cloudberry incubating project.
> >
> > There is a paragraph on the Attic website (see “Process for Leaving the
> > Attic” paragraph), but it doesn’t say much more than what I’ve said above:
> >
> > https://attic.apache.org/
> >
> >
> > So the next steps could be to first ask the Board if a resolution to
> > re-establish MADlib is an option. If they don’t accept that, then discuss
> > with the incubator what the best option is (re-incubating MADlib as it’s on
> > project or as a sub-project of Cloudberry).
> >
> > 2. While this is being discussed, could the project's assets
> > > (repositories, website, JIRA, mailing list archives) be kept in
> > > their current state, so that they can be restored easily if the
> > > board approves?
> >
> >
> > We haven’t yet started doing anything with MADlib. The last step is to ask
> > Infra to make repos & JIRA read-only (which is reversible) and remove
> > releases & mailing lists. You have some time before we start that & we can
> > watch what happens while you work out the best route.
> >
> > Regards
> >
> > Niall
> >
> >
> > > 3. We welcome any feedback or concerns from the Attic team, the
> > > board, and former MADlib PMC members.
> > >
> > > Thank you for your time.
> > >
> > > [1] https://lists.apache.org/thread/zf4opv7omfcrdpr668j5bt9w8mf03m3l
> > > [2] https://github.com/apache/madlib/pull/627
> > > [3] https://lists.apache.org/thread/zdogpfdb6og54pnsohl99ojxkjc5c0zz
> > > [4] https://lists.apache.org/thread/781rspl421vcmclzjdqlstzq8b5vgf4m
> > >
> > > Best,
> > > Dianjin Wang
> > >