Обсуждение: Is zheap on track for PostgreSQL 12.0?
I just read an article about the advantages of zheap over the heap that PostgreSQL currently uses and there were several posts on this list about it being more compact as well as more performant. I am hopeful that it makes it into PostgreSQL 12.0, but it is unclear what the current status of this change is.
Can one of the Postgres devs please chime in on its status for 12.0?
Thanks,
Michael Goldshteyn
On Fri, Feb 1, 2019 at 8:29 PM Michael Goldshteyn <mgold10000@gmail.com> wrote: > > I just read an article about the advantages of zheap over the heap that PostgreSQL currently uses and there were severalposts on this list about it being more compact as well as more performant. I am hopeful that it makes it into PostgreSQL12.0, but it is unclear what the current status of this change is. > > Can one of the Postgres devs please chime in on its status for 12.0? > I don't think there is any chance for zheap to get into PG12.0. However, we have made good progress by rebasing it over pluggable API (thanks to Andres for doing the leg work for same). We have also made quite a few performance improvements in the last few months. Right now, we are doing some stress testing of the core zheap and trying to finish the remaining things like support for logical decoding. Also, we are trying to make progress on the undo patches by reviewing the code which is one of the building blocks for zheap. I request you to help in review/test of zheap if you want to see more progress on the same. -- With Regards, Amit Kapila. EnterpriseDB: http://www.enterprisedb.com
Hopefully someone checked out this academic paper, as the perf gains seem impressive on the surface.. http://madsys.cs.tsinghua.edu.cn/publications/TS2018-liu.pdf DudeTx: Durable Transactions Made Decoupled "This paper presents DudeTx, a crash-consistent durable transaction system that avoids the drawbacks of both undo and redo logging." -- Sent from: http://www.postgresql-archive.org/PostgreSQL-hackers-f1928748.html