Streaming decoding fails with "unexpected table_index_fetch_tuple call during logical decoding" when a relation has a TOASTed conbin (follow-up to BUG #18641)

Поиск
Список
Период
Сортировка
Искать

Streaming decoding fails with "unexpected table_index_fetch_tuple call during logical decoding" when a relation has a TOASTed conbin (follow-up to BUG #18641)

От:
Jiří Kavalík <jiri.kavalik@comgate.cz>
Дата:
Вложения:

Version and platform

PostgreSQL 18.3 (Debian 18.3-1.pgdg13+1) on x86_64-pc-linux-gnu, compiled by gcc (Debian 14.2.0-19) 14.2.0, 64-bit Official postgres:18 Docker image: Debian GNU/Linux 13 (trixie), glibc 2.41, Linux 6.17 x86_64. Seen in production on the same version with pgoutput / built-in logical replication (subscriptions with streaming = parallel). In production, the table with the TOASTed conbin (followed by 34 more pg_constraint rows) is modified by the streamed transaction but is not a member of the subscribed publication. The same holds in the reproducer: with pgoutput and a publication containing only "filler", decoding still fails, because the relcache entry is built for every changed relation.

Non-default settings:

wal_level = logical
logical_decoding_work_mem = 64kB

Steps to reproduce

The attached repro.sh runs these steps with psql against an empty database:

CREATE TABLE t (id int PRIMARY KEY, v text,
CONSTRAINT a_big CHECK (v <> ALL (ARRAY[ <400 md5() literals> ])));
-- the conbin of a_big is stored out of line:
-- pg_column_toast_chunk_id(conbin) IS NOT NULL
CREATE TABLE filler (id int, pad text);
SELECT pg_create_logical_replication_slot('repro', 'test_decoding');
-- session 1, kept open:
BEGIN;
INSERT INTO t VALUES (1, 'y');
INSERT INTO filler SELECT i, repeat('x', 200) FROM generate_series(1, 10000) i;
SELECT pg_sleep(10);
COMMIT;
-- session 2, a new backend, while session 1 is in pg_sleep:
\set VERBOSITY verbose
SELECT count(*) FROM pg_logical_slot_peek_changes('repro', NULL, NULL, 'stream-changes', '1');
-- after session 1 has committed:
SELECT count(*) FROM pg_logical_slot_peek_changes('repro', NULL, NULL, 'stream-changes', '1');

Actual output

Session 2, while session 1 is still open (3 out of 3 runs on a fresh container):

ERROR: XX000: unexpected table_index_fetch_tuple call during logical decoding
LOCATION: table_index_fetch_tuple, tableam.h:1218

The same peek after session 1 has committed: 10106 rows, no error.

The same error occurs with pgoutput (proto_version '4', streaming 'parallel').

Expected output

The in-progress transaction is streamed without error, as it is when the relation has no TOASTed catalog data.

Variations (same procedure, different table definitions)

pg_constraint rows of the table, in conname orderResult
a_big (TOASTed), b_small, t_id_not_null, t_pkeyERROR
a_big (TOASTed), t_only_id_not_null, t_only_pkeyERROR
a_small, t_last_id_not_null, t_last_pkey, z_big (TOASTed)no error
no TOASTed conbinno error

The error occurs only when the TOASTed row is not the last one the scan returns. Because NOT NULL constraints are pg_constraint rows in PG18, most tables have rows after any given CHECK constraint.

Backtrace

Captured with backtrace_functions = 'table_index_fetch_tuple'. The server has no debug symbols, so static functions show as offsets:

(+0xebb28)
index_getnext_slot+0x45
systable_getnext+0x38
(+0x66feae)
RelationIdGetRelation+0x8d
(+0x4a304e)
(+0x4a42b5)
ReorderBufferQueueChange+0x301
heap_decode+0x1df
LogicalDecodingProcessRecord+0x76
(+0x49afa5)
pg_logical_slot_peek_changes+0x11

Possible cause (a suggestion only)

bsysscan is a plain bool, set and cleared in systable_{begin,end}scan[_ordered]. Since 8175a7d11 (the fix for BUG #18641), the TOAST fetch that detoasts conbin inside the outer pg_constraint scan clears bsysscan on its end-scan. The outer scan's next systable_getnext() then reaches table_index_fetch_tuple() while CheckXidAlive is still valid. That would also explain why a TOASTed row in the last position does not fail. Restoring the previous value of bsysscan, or counting nesting depth, instead of clearing it might fix this.

Re: Streaming decoding fails with "unexpected table_index_fetch_tuple call during logical decoding" when a relation has a TOASTed conbin (follow-up to BUG #18641)

От:
"Hayato Kuroda (Fujitsu)" <kuroda.hayato@fujitsu.com>
Дата:
Вложения:
Hi,

This is the reply for [1]. My mailer could not receive the original post due to
the company's policy, so I will put as the normal post. Sorry for inconvenience.

I confirmed this could happen on PG18, PG17 and PG14. Not tested, but expecting
for PG15 and 166 as well. This could not happen on PG19/HEAD because the
elog(ERROR) was removed by 87f7b824f20, but possible for all branches.

I think your analysis is correct. bsysscan tries to indicate that whether we're
scanning a system table, which was turned on at systable_beginscan* and turned
off at systable_endscan*. But if the systable scan is nested (i.e., pg_constraint.conbin),
the flag can be wrong reset. In PG18- the state is checked for every getnextslot,
which raised the ERROR. In PG19+ the check is unified at the beginning thus the
ERROR does not happen, but I guess the flag can be still wrong.

One idea is to track the depth of scans. Attached patch is for PG18, and I tried not
to modify the header as much as possible. It also had a test code based on your
reproducer. Can you see it's same as your expectation?

[1]: https://www.postgresql.org/message-id/CAF7a2M-OF%2BTjYUBvkSV2oyZNYuPm%3DEqTApcMDOC%2BaevBFPuWrg%40mail.gmail.com

Best regards,
Hayato Kuroda
FUJITSU LIMITED

Re: Streaming decoding fails with "unexpected table_index_fetch_tuple call during logical decoding" when a relation has a TOASTed conbin (follow-up to BUG #18641)

От:
Jiří Kavalík <jiri.kavalik@comgate.cz>
Дата:
Hi Kuroda-san,

Thank you for the quick analysis and the patch.

> Can you see it's same as your expectation?

Yes. I tested v1 on REL_18_STABLE (37bbf5bba0, 18.6 plus later commits), with
two builds from the same tree differing only by the patch, both configured
with --enable-debug --enable-cassert.

In each case below, a fresh backend peeks at a slot while a transaction that
first inserts one row into the test table, then 10000 rows into a filler
table, is still open (logical_decoding_work_mem = 64kB, CHECKPOINT before the
peek):

1. TOASTed conbin first, three more constraints after it:
   - test_decoding: unpatched ERROR, patched OK
   - pgoutput, table published: unpatched ERROR, patched OK
   - pgoutput, table NOT published: unpatched ERROR, patched OK

2. TOASTed conbin first, only the NOT NULL and PK constraints after it:
   - test_decoding: unpatched ERROR, patched OK
   - pgoutput, table published: unpatched ERROR, patched OK
   - pgoutput, table NOT published: unpatched ERROR, patched OK

3. TOASTed conbin last: OK on both builds, both plugins.

4. No TOASTed conbin: OK on both builds, both plugins.

"OK" means the whole transaction was decoded (10098 rows from test_decoding,
10100 from pgoutput). There were no assertion failures in either server log.

Your new stream.sql test fails on the unpatched build with the same ERROR and
passes with the patch. With the patch, contrib/test_decoding "make check"
(19 regression + 13 isolation tests) and the core "make check" (231 tests)
pass.

The "table NOT published" rows may be worth noting: the relation that hit the
error in our production case is not in the subscribed publication, it is only
modified by the streamed transaction. The relcache entry is still built, so
the bug is reachable for any table the transaction touches.

One question while reading the patch, not a problem I could trigger: the depth
is incremented in systable_beginscan* and decremented in systable_endscan* only
if CheckXidAlive is valid, and that is evaluated separately at each end. If
CheckXidAlive changed while a scan was open, the counter would be off by one.
I could not find a path where that happens. Error paths look fine, since
AbortTransaction/AbortSubTransaction call ResetLogicalStreamingState(). If a
SysScanDesc field is acceptable despite the header concern, remembering in the
scan whether it was counted would make the pairing explicit.

Regards,
Jiří Kavalík


pá 2. 10. 2026 v 9:33 odesílatel Hayato Kuroda (Fujitsu) <kuroda.hayato@fujitsu.com> napsal:
Hi,

This is the reply for [1]. My mailer could not receive the original post due to
the company's policy, so I will put as the normal post. Sorry for inconvenience.

I confirmed this could happen on PG18, PG17 and PG14. Not tested, but expecting
for PG15 and 166 as well. This could not happen on PG19/HEAD because the
elog(ERROR) was removed by 87f7b824f20, but possible for all branches.

I think your analysis is correct. bsysscan tries to indicate that whether we're
scanning a system table, which was turned on at systable_beginscan* and turned
off at systable_endscan*. But if the systable scan is nested (i.e., pg_constraint.conbin),
the flag can be wrong reset. In PG18- the state is checked for every getnextslot,
which raised the ERROR. In PG19+ the check is unified at the beginning thus the
ERROR does not happen, but I guess the flag can be still wrong.

One idea is to track the depth of scans. Attached patch is for PG18, and I tried not
to modify the header as much as possible. It also had a test code based on your
reproducer. Can you see it's same as your expectation?

[1]: https://www.postgresql.org/message-id/CAF7a2M-OF%2BTjYUBvkSV2oyZNYuPm%3DEqTApcMDOC%2BaevBFPuWrg%40mail.gmail.com

Best regards,
Hayato Kuroda
FUJITSU LIMITED



--
S pozdravem
Jiří Kavalík
Comgate a.s.
Gočárova třída 1754/48b, 500 02 Hradec Králové

RE: Streaming decoding fails with "unexpected table_index_fetch_tuple call during logical decoding" when a relation has a TOASTed conbin (follow-up to BUG #18641)

От:
"Hayato Kuroda (Fujitsu)" <kuroda.hayato@fujitsu.com>
Дата:
Hi Jiří,

Thanks for the test. I understood that no issues were found for now.

> One question while reading the patch, not a problem I could trigger: the depth
> is incremented in systable_beginscan* and decremented in systable_endscan* only
> if CheckXidAlive is valid, and that is evaluated separately at each end. If
> CheckXidAlive changed while a scan was open, the counter would be off by one.
> I could not find a path where that happens. Error paths look fine, since
> AbortTransaction/AbortSubTransaction call ResetLogicalStreamingState(). If a
> SysScanDesc field is acceptable despite the header concern, remembering in the
> scan whether it was counted would make the pairing explicit.

Hmm, I'm not excited to modify the exposed data structure yet, unless there is a
real issue. Per my analysis, SysScanDescData only contains pointers (8 bytes),
it does not have any paddings. This meant we need to modify a size of the data
structure, it might cause failures somewhere.
Also, the existing code has the same possibility while turning on/off bsysscan,
right? So I feel it's already accepted.
(Of course, we must fix if it causes a real failure)

Best regards,
Hayato Kuroda
FUJITSU LIMITED

Re: Streaming decoding fails with "unexpected table_index_fetch_tuple call during logical decoding" when a relation has a TOASTed conbin (follow-up to BUG #18641)

От:
Jiří Kavalík <jiri.kavalik@comgate.cz>
Дата:
Hi Kuroda-san,

> Hmm, I'm not excited to modify the exposed data structure yet, unless there is a
> real issue.

Agreed. It was only a readability suggestion, not a fix for a real failure.
I checked again. CheckXidAlive is set only at the start of each change in
ReorderBufferProcessTXN(), before that change's catalog access. It is cleared
only after the change loop, or on abort, where ResetLogicalStreamingState()
also resets the new depth counter. So it does not change while a scan is open,
except on the error path, which v1 already handles. v1 looks complete to me.

Thanks for the patch.

Regards,
Jiří Kavalík

út 6. 10. 2026 v 11:22 odesílatel Hayato Kuroda (Fujitsu) <kuroda.hayato@fujitsu.com> napsal:
Hi Jiří,

Thanks for the test. I understood that no issues were found for now.

> One question while reading the patch, not a problem I could trigger: the depth
> is incremented in systable_beginscan* and decremented in systable_endscan* only
> if CheckXidAlive is valid, and that is evaluated separately at each end. If
> CheckXidAlive changed while a scan was open, the counter would be off by one.
> I could not find a path where that happens. Error paths look fine, since
> AbortTransaction/AbortSubTransaction call ResetLogicalStreamingState(). If a
> SysScanDesc field is acceptable despite the header concern, remembering in the
> scan whether it was counted would make the pairing explicit.

Hmm, I'm not excited to modify the exposed data structure yet, unless there is a
real issue. Per my analysis, SysScanDescData only contains pointers (8 bytes),
it does not have any paddings. This meant we need to modify a size of the data
structure, it might cause failures somewhere.
Also, the existing code has the same possibility while turning on/off bsysscan,
right? So I feel it's already accepted.
(Of course, we must fix if it causes a real failure)

Best regards,
Hayato Kuroda
FUJITSU LIMITED



--
S pozdravem
Jiří Kavalík
Comgate a.s.
Gočárova třída 1754/48b, 500 02 Hradec Králové

Re: Streaming decoding fails with "unexpected table_index_fetch_tuple call during logical decoding" when a relation has a TOASTed conbin (follow-up to BUG #18641)

От:
Zhijie Hou <houzhijie22@gmail.com>
Дата:
Hi,

On Fri, Oct 2, 2026 at 3:33 PM Hayato Kuroda (Fujitsu)
 wrote:
>
> Hi,
>
> This is the reply for [1]. My mailer could not receive the original post due to
> the company's policy, so I will put as the normal post. Sorry for inconvenience.
>
> I confirmed this could happen on PG18, PG17 and PG14. Not tested, but expecting
> for PG15 and 166 as well. This could not happen on PG19/HEAD because the
> elog(ERROR) was removed by 87f7b824f20, but possible for all branches.
>
> I think your analysis is correct. bsysscan tries to indicate that whether we're
> scanning a system table, which was turned on at systable_beginscan* and turned
> off at systable_endscan*. But if the systable scan is nested (i.e., pg_constraint.conbin),
> the flag can be wrong reset. In PG18- the state is checked for every getnextslot,
> which raised the ERROR. In PG19+ the check is unified at the beginning thus the
> ERROR does not happen, but I guess the flag can be still wrong.
>
> One idea is to track the depth of scans. Attached patch is for PG18, and I tried not
> to modify the header as much as possible. It also had a test code based on your
> reproducer. Can you see it's same as your expectation?

I think using a counter to fix this is OK. But I'm wondering whether it's
necessary to keep both the boolean and the counter, could we merge them into a
single parameter to simplify things?

Best Regards,
Zhijie Hou


RE: Streaming decoding fails with "unexpected table_index_fetch_tuple call during logical decoding" when a relation has a TOASTed conbin (follow-up to BUG #18641)

От:
"Hayato Kuroda (Fujitsu)" <kuroda.hayato@fujitsu.com>
Дата:
Вложения:
Dear Hou,

Thanks for the review!

> I think using a counter to fix this is OK. But I'm wondering whether it's
> necessary to keep both the boolean and the counter, could we merge them into
> a
> single parameter to simplify things?

Actually it's intentional because the initially attached patch was for PG18.
bsysscan was exported and functions in tableam.h referred it. I did not want to
modify the header as much as possible thus they were retained.

For HEAD, yes, the boolean can be removed from everywhere, see attached.
I will create patched for back branches when they are stabilized.

Best regards,
Hayato Kuroda
FUJITSU LIMITED

Re: Streaming decoding fails with "unexpected table_index_fetch_tuple call during logical decoding" when a relation has a TOASTed conbin (follow-up to BUG #18641)

От:
Jiří Kavalík <jiri.kavalik@comgate.cz>
Дата:
Hi Kuroda-san,

> For HEAD, yes, the boolean can be removed from everywhere, see attached.

One small thing in v2-HEAD: the definition in xact.c is left behind.
With the patch applied on master (061065e), "git grep -w bsysscan" still finds

    src/backend/access/transam/xact.c:103:bool bsysscan = false;

and since the extern is gone from xact.h, gcc 14 warns:

    xact.c:103:17: warning: no previous declaration for 'bsysscan' [-Wmissing-variable-declarations]

Removing that line should be enough. The comment above it in xact.c may also
want a word about the counter.

Regards,
Jiří Kavalík


čt 8. 10. 2026 v 4:36 odesílatel Hayato Kuroda (Fujitsu) <kuroda.hayato@fujitsu.com> napsal:
Dear Hou,

Thanks for the review!

> I think using a counter to fix this is OK. But I'm wondering whether it's
> necessary to keep both the boolean and the counter, could we merge them into
> a
> single parameter to simplify things?

Actually it's intentional because the initially attached patch was for PG18.
bsysscan was exported and functions in tableam.h referred it. I did not want to
modify the header as much as possible thus they were retained.

For HEAD, yes, the boolean can be removed from everywhere, see attached.
I will create patched for back branches when they are stabilized.

Best regards,
Hayato Kuroda
FUJITSU LIMITED



--
S pozdravem
Jiří Kavalík
Comgate a.s.
Gočárova třída 1754/48b, 500 02 Hradec Králové

RE: Streaming decoding fails with "unexpected table_index_fetch_tuple call during logical decoding" when a relation has a TOASTed conbin (follow-up to BUG #18641)

От:
"Hayato Kuroda (Fujitsu)" <kuroda.hayato@fujitsu.com>
Дата:
Вложения:
Dear Jiří

Good catch. I accidentally removed the meson option " --werror" so that
I could not recognize. PSA fixed version.

Best regards,
Hayato Kuroda
FUJITSU LIMITED

RE: Streaming decoding fails with "unexpected table_index_fetch_tuple call during logical decoding" when a relation has a TOASTed conbin (follow-up to BUG #18641)

От:
"Hayato Kuroda (Fujitsu)" <kuroda.hayato@fujitsu.com>
Дата:
Вложения:
Hi members,

I made a patch set for supported branches. I confirmed the error can happen
and the patch can address the issue on PG14-PG18.

I refactored the test code for the readability. Wondering PG19/HEAD should contain
the test, because the reported ERROR could not happen in the first place.

Best regards,
Hayato Kuroda
FUJITSU LIMITED

Re: Streaming decoding fails with "unexpected table_index_fetch_tuple call during logical decoding" when a relation has a TOASTed conbin (follow-up to BUG #18641)

От:
Peter Smith <smithpb2250@gmail.com>
Дата:
On Tue, Oct 27, 2026 at 12:25 AM Hayato Kuroda (Fujitsu)
 wrote:
>
> Hi members,
>
> I made a patch set for supported branches. I confirmed the error can happen
> and the patch can address the issue on PG14-PG18.
>
> I refactored the test code for the readability. Wondering PG19/HEAD should contain
> the test, because the reported ERROR could not happen in the first place.
>

Hi Kuroda-San,

Since you are refactoring this flag anyway, would it be better to try
to eliminate the global-variable completely? e.g. expose the same
information by some accessor function.

-extern PGDLLIMPORT bool bsysscan;
+extern PGDLLIMPORT int sysscan_depth;

======
Kind Regards,
Peter Smith.
Fujitsu Australia.


RE: Streaming decoding fails with "unexpected table_index_fetch_tuple call during logical decoding" when a relation has a TOASTed conbin (follow-up to BUG #18641)

От:
"Hayato Kuroda (Fujitsu)" <kuroda.hayato@fujitsu.com>
Дата:
Вложения:
Hi Peter,

> Since you are refactoring this flag anyway, would it be better to try
> to eliminate the global-variable completely? e.g. expose the same
> information by some accessor function.
> 
> -extern PGDLLIMPORT bool bsysscan;
> +extern PGDLLIMPORT int sysscan_depth;

Good point, it can be removed by providing few functions. v5 did the same.
We need to keep bsysscan anyway for back branches, so the patch shape for them
is slightly different, see v5-PG18.

Best regards,
Hayato Kuroda
FUJITSU LIMITED

Re: Streaming decoding fails with "unexpected table_index_fetch_tuple call during logical decoding" when a relation has a TOASTed conbin (follow-up to BUG #18641)

От:
Chao Li <li.evan.chao@gmail.com>
Дата:


> On Oct 27, 2026, at 14:42, Hayato Kuroda (Fujitsu)  wrote:
> 
> Hi Peter,
> 
>> Since you are refactoring this flag anyway, would it be better to try
>> to eliminate the global-variable completely? e.g. expose the same
>> information by some accessor function.
>> 
>> -extern PGDLLIMPORT bool bsysscan;
>> +extern PGDLLIMPORT int sysscan_depth;
> 
> Good point, it can be removed by providing few functions. v5 did the same.
> We need to keep bsysscan anyway for back branches, so the patch shape for them
> is slightly different, see v5-PG18.
> 
> Best regards,
> Hayato Kuroda
> FUJITSU LIMITED
> 
> 

Hi Hayato,

I agree with the solution. However, I have a minor concern about the naming. BeginSystemTableScan() and EndSystemTableScan() sound like functions that actually start and end system table scans, whereas they only track sysscan_depth. How about renaming them to TrackSystemTableScanStart() and TrackSystemTableScanEnd() to better reflect their purpose?

Best regards,
--
Chao Li (Evan)
HighGo Software Co., Ltd.
https://www.highgo.com/






RE: Streaming decoding fails with "unexpected table_index_fetch_tuple call during logical decoding" when a relation has a TOASTed conbin (follow-up to BUG #18641)

От:
"Hayato Kuroda (Fujitsu)" <kuroda.hayato@fujitsu.com>
Дата:
Вложения:
Hi Chao,

> I agree with the solution. However, I have a minor concern about the naming.
> BeginSystemTableScan() and EndSystemTableScan() sound like functions that
> actually start and end system table scans, whereas they only track sysscan_depth.
> How about renaming them to TrackSystemTableScanStart() and
> TrackSystemTableScanEnd() to better reflect their purpose?

+1, PSA new version.

Best regards,
Hayato Kuroda
FUJITSU LIMITED

Re: Streaming decoding fails with "unexpected table_index_fetch_tuple call during logical decoding" when a relation has a TOASTed conbin (follow-up to BUG #18641)

От:
Shlok Kyal <shlok.kyal.oss@gmail.com>
Дата:
On Thu, 29 Oct 2026 at 10:22, Hayato Kuroda (Fujitsu)
 wrote:
>
> Hi Chao,
>
> > I agree with the solution. However, I have a minor concern about the naming.
> > BeginSystemTableScan() and EndSystemTableScan() sound like functions that
> > actually start and end system table scans, whereas they only track sysscan_depth.
> > How about renaming them to TrackSystemTableScanStart() and
> > TrackSystemTableScanEnd() to better reflect their purpose?
>
> +1, PSA new version.
>
Hi Kuroda-san,

I have reviewed and tested the patches.
I am able to reproduce the issue in PG14 to PG18 using the test added
in stream.sql and confirm that the corresponding patches address the
issue.
Patches is applying cleanly on all branches and passed 'make check-world'

I have some minor comments:
1. For all patches: Most of the comments for function in the file
'xact.c' is like of format:
/*
 * Function name
 *
 * Function Description
 */
Maybe we should follow the same for the functions added by these patches?

2. For patches for PG14 to PG18: In logical.c we can remove:
+#include "access/genam.h"
 #include "access/xact.h"

In logical.c we only introduced "ResetSystemTableScanState();" and
this function which was declared 'access/xact.h'
logical.c already includes access/xact.h

3. For patches for PG 18 and PG 17:  I noticed a unrelated line deletion change:
 #include "utils/snapmgr.h"

-
 /* ----------------------------------------------------------------

Was it intentional?

4. For patches for PG 14 and PG 15:
+    ALTER TABLE nested_sys_scan_test ALTER COLUMN checked_value set
storage external;
I think we should capitalize 'set storage external;'

Thanks,
Shlok Kyal


RE: Streaming decoding fails with "unexpected table_index_fetch_tuple call during logical decoding" when a relation has a TOASTed conbin (follow-up to BUG #18641)

От:
"Hayato Kuroda (Fujitsu)" <kuroda.hayato@fujitsu.com>
Дата:
Вложения:
Hi Shlok,

Thanks for the review!

> 1. For all patches: Most of the comments for function in the file
> 'xact.c' is like of format:
> /*
>  * Function name
>  *
>  * Function Description
>  */
> Maybe we should follow the same for the functions added by these patches?

I do not have strong opinions for it, but added accordingly.

> 2. For patches for PG14 to PG18: In logical.c we can remove:
> +#include "access/genam.h"
>  #include "access/xact.h"
> 
> In logical.c we only introduced "ResetSystemTableScanState();" and
> this function which was declared 'access/xact.h'
> logical.c already includes access/xact.h

Fixed. It was included because functions are initially declared in genam.h.

> 3. For patches for PG 18 and PG 17:  I noticed a unrelated line deletion change:
>  #include "utils/snapmgr.h"
> 
> -
>  /* ----------------------------------------------------------------
> 
> Was it intentional?

No, removed.

> 4. For patches for PG 14 and PG 15:
> +    ALTER TABLE nested_sys_scan_test ALTER COLUMN checked_value set
> storage external;
> I think we should capitalize 'set storage external;'

+1. Fixed.

PSA new patch set.

Best regards,
Hayato Kuroda
FUJITSU LIMITED

Re: Streaming decoding fails with "unexpected table_index_fetch_tuple call during logical decoding" when a relation has a TOASTed conbin (follow-up to BUG #18641)

От:
Shlok Kyal <shlok.kyal.oss@gmail.com>
Дата:
>
> PSA new patch set.
>
Hi Kuroda-san,

I have reviewed the patches and it looks good to me.

Thanks,
Shlok Kyal


RE: Streaming decoding fails with "unexpected table_index_fetch_tuple call during logical decoding" when a relation has a TOASTed conbin (follow-up to BUG #18641)

От:
"Hayato Kuroda (Fujitsu)" <kuroda.hayato@fujitsu.com>
Дата:
Вложения:
Hi,

Since PG19 is now releasing, it's not a good idea to modify the header.
An exported variable bsysscan is retained in this version.
Patches for other versions are unchanged.

Best regards,
Hayato Kuroda
FUJITSU LIMITED

FAQ