There has been some interest in continuing to use Pulsar SQL in the future. To facilitate this, I have taken the following actions: - Extracted the pulsar-sql files from branch-3.0 to a new repository: https://github.com/apache/pulsar-sql - This new repository allows interested volunteers to make contributions, which has been a barrier until now. - The README file in https://github.com/apache/pulsar-sql explains how the pulsar-sql files were extracted with commit history to the new repository.
Please note: - The new repository currently lacks a build system and CI setup. - Setting up the build and CI will be the first task for volunteers interested in maintaining this component. We welcome contributions from the community to help maintain and improve Pulsar SQL going forward. In the meantime, current Pulsar SQL users can continue to use Pulsar 3.0.x LTS, which is supported until May 2025. There are more details in the GitHub Discussions answer, https://github.com/apache/pulsar/discussions/23435 -Lari On 2023/12/22 17:09:19 Matteo Merli wrote: > I want to start a discussion regarding the removal of all the code related > to the Trino (PrestoDB) plugin from the Pulsar main repository. > > This topic was already discussed and approved long time ago in PIP-62 ( > https://github.com/apache/pulsar/wiki/PIP-62%3A-Move-connectors%2C-adapters-and-Pulsar-Presto-to-separate-repositories > ) > > The main reasons for not having Presto plugin as part of the main > distribution of Pulsar were (and still are valid): > > 1. We need to ship the entire Presto runtime which is ~400 MB. This makes > our tgz and Docker images huge > 3. There is no strict need for this component to be in the same > distribution / image: it could easily be provided in a different release > tgz or Docker image > > Though I think that since then it became more clear that the current state > of this plugin has been stagnating over the years. > > 1. There are not many active users of Pulsar-SQL component (I'd be very > happy to be contradicted here) > 2. The plugin code has not been improved in a long time > 3. There are several open security issues (actually, almost the totality of > current dependencies issues are today coming from Trino). > > My suggestion would be that, if there is any volunteer willing to pick this > plugin up and maintain it in a separate repository (within the Apache > Pulsar project) and with a separate release schedule, we should go ahead > and move it. > If there are no volunteers, we should just remove it as it is. If later on > we want to revive it, we can always import the code from the last commit. > > Thoughts? > > > -- > Matteo Merli > <mme...@apache.org> >
