From: | "Zhijie Hou (Fujitsu)" <houzj(dot)fnst(at)fujitsu(dot)com> |
---|---|
To: | Amit Kapila <amit(dot)kapila16(at)gmail(dot)com> |
Cc: | "Hayato Kuroda (Fujitsu)" <kuroda(dot)hayato(at)fujitsu(dot)com>, shveta malik <shveta(dot)malik(at)gmail(dot)com>, Nisha Moond <nisha(dot)moond412(at)gmail(dot)com>, pgsql-hackers <pgsql-hackers(at)postgresql(dot)org>, Dilip Kumar <dilipbalaut(at)gmail(dot)com>, Jan Wieck <jan(at)wi3ck(dot)info>, Ashutosh Bapat <ashutosh(dot)bapat(dot)oss(at)gmail(dot)com> |
Subject: | RE: Conflict detection and logging in logical replication |
Date: | 2024-08-14 02:35:34 |
Message-ID: | OS0PR01MB5716DFA90DEF51ADA8E8E53594872@OS0PR01MB5716.jpnprd01.prod.outlook.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On Tuesday, August 13, 2024 7:04 PM Amit Kapila <amit(dot)kapila16(at)gmail(dot)com> wrote:
>
> On Tue, Aug 13, 2024 at 10:09 AM Zhijie Hou (Fujitsu)
> <houzj(dot)fnst(at)fujitsu(dot)com> wrote:
> >
> > Here is V13 patch set which addressed above comments.
> >
>
> 1.
> +ReportApplyConflict(int elevel, ConflictType type, EState *estate,
> +ResultRelInfo *relinfo,
>
> The change looks better but it would still be better to keep elevel and type after
> relinfo. The same applies to other places as well.
Changed.
>
> 2.
> + * The caller should ensure that the index with the OID 'indexoid' is locked.
> + *
> + * Refer to errdetail_apply_conflict for the content that will be
> +included in
> + * the DETAIL line.
> + */
> +void
> +ReportApplyConflict(int elevel, ConflictType type, EState *estate,
>
> Is it possible to add an assert to ensure that the index is locked by the caller?
Added.
>
> 3.
> +static char *
> +build_tuple_value_details(EState *estate, ResultRelInfo *relinfo,
> + TupleTableSlot *searchslot,
> + TupleTableSlot *localslot,
> + TupleTableSlot *remoteslot,
> + Oid indexoid)
> {
> ...
> ...
> + /*
> + * If 'searchslot' is NULL and 'indexoid' is valid, it indicates that
> + we
> + * are reporting the unique constraint violation conflict, in which
> + case
> + * the conflicting key values will be reported.
> + */
> + if (OidIsValid(indexoid) && !searchslot) {
> ...
> ...
> }
>
> This indirect way of inferencing constraint violation looks fragile.
> The caller should pass the required information explicitly and then you can
> have the required assertions here.
>
> Apart from the above, I have made quite a few changes in the code comments
> and LOG messages in the attached.
Thanks. I have addressed above comments and merged the changes.
Here is the V14 patch.
Best Regards,
Hou zj
Attachment | Content-Type | Size |
---|---|---|
v14-0001-Detect-and-log-conflicts-in-logical-replication.patch | application/octet-stream | 63.1 KB |
From | Date | Subject | |
---|---|---|---|
Next Message | Imseih (AWS), Sami | 2024-08-14 02:40:48 | Re: query_id, pg_stat_activity, extended query protocol |
Previous Message | Thomas Munro | 2024-08-14 02:31:20 | Re: Remaining dependency on setlocale() |