Too rigorous assert in reorderbuffer.c

From: Arseny Sher <a(dot)sher(at)postgrespro(dot)ru>
To: "pgsql-hackers" <pgsql-hackers(at)postgresql(dot)org>, Alexander Lakhin <a(dot)lakhin(at)postgrespro(dot)ru>
Subject: Too rigorous assert in reorderbuffer.c
Date: 2019-01-31 06:21:59
Message-ID: 874l9p8hyw.fsf@ars-thinkpad
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-hackers


My colleague Alexander Lakhin has noticed an assertion failure in
reorderbuffer.c:1330. Here is a simple snippet reproducing it:

SELECT 'init' FROM pg_create_logical_replication_slot('regression_slot', 'test_decoding');

create table t(k int);
savepoint a;
alter table t alter column k type text;
rollback to savepoint a;
alter table t alter column k type bigint;

SELECT data FROM pg_logical_slot_get_changes('regression_slot', NULL, NULL, 'include-xids', '0', 'skip-empty-xacts', '1');

It is indeed too opinionated since cmax of a tuple is not stable; it can
be rewritten if subxact who tried to delete it later aborts (analogy
also holds for xmax). Attached patch removes it. While here, I had also
considered worthwhile to add a test involving DDL in aborted subxact as
it is interesting anyway and wasn't covered before.

Attachment Content-Type Size
0001-Remove-assertion-in-reorderbuffer-that-cmax-is-stabl.patch text/x-diff 3.8 KB


Browse pgsql-hackers by date

  From Date Subject
Next Message Kyotaro HORIGUCHI 2019-01-31 06:31:53 Re: Index Skip Scan
Previous Message Shawn Debnath 2019-01-31 05:59:38 Re: Refactoring the checkpointer's fsync request queue