Re: Improve eviction algorithm in ReorderBuffer
Masahiko Sawada <sawada.mshk@gmail.com>
From: Masahiko Sawada <sawada.mshk@gmail.com>
To: Amit Kapila <amit.kapila16@gmail.com>
Cc: Dilip Kumar <dilipbalaut@gmail.com>, PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2023-12-15T12:18:55Z
Lists: pgsql-hackers
On Fri, Dec 15, 2023 at 2:59 PM Masahiko Sawada <sawada.mshk@gmail.com> wrote: > > On Fri, Dec 15, 2023 at 12:37 PM Amit Kapila <amit.kapila16@gmail.com> wrote: > > > > On Wed, Dec 13, 2023 at 6:01 AM Masahiko Sawada <sawada.mshk@gmail.com> wrote: > > > > > > On Tue, Dec 12, 2023 at 1:33 PM Dilip Kumar <dilipbalaut@gmail.com> wrote: > > > > > > > > On Tue, Dec 12, 2023 at 9:01 AM Masahiko Sawada <sawada.mshk@gmail.com> wrote: > > > > > > > > > > I've heard the report internally that replication lag became huge when > > > > > decoding transactions each consisting of 500k sub transactions. Note > > > > > that ReorderBufferLargetstTXN() is used only in non-streaming mode. > > > > > > > > > Can't you suggest them to use streaming mode to avoid this problem or > > do you see some problem with that? > > Yeah, that's one option. But I can suggest > Sorry, it was still in the middle of editing. Yeah, that's one option. But since there is a trade-off I cannot suggest using streaming mode for every user. Also, the logical replication client (e.g. third party tool receiving logical change set) might not support the streaming mode yet. Regards, -- Masahiko Sawada Amazon Web Services: https://aws.amazon.com
Commits
-
Revert indexed and enlargable binary heap implementation.
- 810f64a01567 17.0 landed
-
Replace binaryheap + index with pairingheap in reorderbuffer.c
- efb8acc0d058 17.0 landed
-
Improve eviction algorithm in ReorderBuffer using max-heap for many subtransactions.
- 5bec1d6bc5e3 17.0 landed
-
Add functions to binaryheap for efficient key removal and update.
- b84050864415 17.0 landed
-
Make binaryheap enlargeable.
- bcb14f4abca0 17.0 landed