Hey Murali, Sorry for not replying earlier. Thanks for the interest in the idea. I would look at a few things to make a solid proposal for this idea. Basically it should be able to connect to AsterixDB, read the Metadata Dataverse, and from that create a set of DDL statements that would generate the rows that would exist in the Metadata. So you should first get a good understanding of how the Metadata is currently serialized from DDL statements, deserialized to be used by the compiler and other parts of the system, and so on.
If there are any bugs,quirks or questions you find in your explorations of asterix-metadata, TupleTranslator(s), or the general workflow of the code around this area, it would be great to engage with the community here about that as a first step. The code in this area is some of the oldest, due to the necessity of needing a catalog to bootstrap the system, and therefore some of the most verbose. That way, hopefully, we can set you on the right path. Any other bugs you might find here or in JIRA that seem easy to fix would also be a good first step. - Ian On Tue, Mar 24, 2026 at 3:49 PM Murali Udutha <[email protected]> wrote: > > Hi AsterixDB Team, > > I am following up on my previous email regarding the Backup/Restore Utility > (ASTERIXDB-3697) for GSoC 2026. I am currently a Master's student at Kiel > University and am very excited about this project. > > I have been studying the LSM-tree storage engine concepts and am finalizing > my technical proposal. Could you please let me know if my Jira account > request (umurali) has been processed, or if there is any other step I > should take to engage with the community before the March 31st deadline? > > Best regards, > > Murali Udutha
