Re: Minimal logical decoding on standbys

Robert Haas <robertmhaas@gmail.com>

From: Robert Haas <robertmhaas@gmail.com>
To: Andres Freund <andres@anarazel.de>
Cc: "Drouvot, Bertrand" <bertranddrouvot.pg@gmail.com>, Alvaro Herrera <alvherre@2ndquadrant.com>, Ibrar Ahmed <ibrar.ahmad@gmail.com>, Amit Khandekar <amitdkhan.pg@gmail.com>, fabriziomello@gmail.com, tushar <tushar.ahuja@enterprisedb.com>, Rahila Syed <rahila.syed@2ndquadrant.com>, pgsql-hackers <pgsql-hackers@postgresql.org>
Date: 2022-12-20T20:39:26Z
Lists: pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →
  1. Reduce the log level in 035_standby_logical_decoding.pl.

  2. 035_standby_logical_decoding: Add missing waits for replication

  3. For cascading replication, wake physical and logical walsenders separately

  4. Handle logical slot conflicts on standby

  5. Support invalidating replication slots due to horizon and wal_level

  6. Prevent use of invalidated logical slot in CreateDecodingContext()

  7. Replace replication slot's invalidated_at LSN with an enum

  8. Pass down table relation into more index relation functions

  9. Assert only valid flag bits are passed to visibilitymap_set()

  10. Remove unused _bt_delitems_delete() argument.

  11. Add xl_btree_delete optimization.

Attachments

On Tue, Dec 20, 2022 at 1:25 PM Robert Haas <robertmhaas@gmail.com> wrote:
> I'm going to go through all the record types one by one before
> commenting further.

OK, so xl_hash_vacuum_one_page, at least, is a live issue. To reproduce:

./configure <whatever your usual options are>
echo 'COPT=-fsanitize=alignment -fno-sanitize-recover=all' > src/Makefile.custom
make -j8
make install
initdb
postgres

Then in another window:

pg_basebackup -D pgstandby -R
# edit postgresql.conf, change port number
postgres -D pgstandby

Then in a third window, using the attached files:

pgbench -i -s 10
psql -f kpt_hash_setup.sql
pgbench -T 10 -c 4 -j 4 -f kpt_hash_pgbench.sql

With the patch, the standby falls over:

bufpage.c:1194:31: runtime error: load of misaligned address
0x7fa62f05d119 for type 'OffsetNumber' (aka 'unsigned short'), which
requires 2 byte alignment
0x7fa62f05d119: note: pointer points here
 00 00 00  00 e5 00 8f 00 00 00 00  87 00 ab 20 20 20 20 20  20 20 20
20 20 20 20 20  20 20 20 20 20
              ^

Without the patch, this doesn't occur.

I think this might be the only WAL record type where there's a
problem, but I haven't fully confirmed that yet.

-- 
Robert Haas
EDB: http://www.enterprisedb.com