From: | Shlok Kyal <shlok(dot)kyal(dot)oss(at)gmail(dot)com> |
---|---|
To: | emre(at)hasegeli(dot)com |
Cc: | pgsql-hackers(at)postgresql(dot)org |
Subject: | Re: Object identifier types in logical replication binary mode |
Date: | 2024-12-24 09:15:32 |
Message-ID: | CANhcyEX95YVu_nJx2KecY-ramO+zUyNP59fakqui79f4-cDrBQ@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Hi,
On Thu, 14 Nov 2024 at 21:22, Emre Hasegeli <emre(at)hasegeli(dot)com> wrote:
>
> I encountered a problem with logical replication.
>
> The object identifier types, regclass, regproc, regtype, etc. are
> transferred as an oid in the binary mode. However, the subscriber
> expects them as text which makes sense because it cannot do anything
> with the oids. I am getting "invalid byte sequence for encoding
> "UTF8": 0x00" when I try this.
>
> I think the publisher should not transfer these in binary just like
> the data types from extensions. Any opinions?
I tried to test the scenario described. Steps I followed:
1. Create a logical replication setup on table t1 with some initial data.
2. Columns in table t1 are of types regclass, regproc and regtype
3. Now, I inserted some records in the table t1 and also did some
updates on the table t1.
I tested with both 'binary = true' and 'binary = false' option while
creating a subscription. For me replication is working fine and I am
not getting any errors in both the cases.
I have also attached the test script.
Am I trying the correct steps? Can you share a reproducible test case?
Also, which version are you facing this issue?
Have you set any GUC parameters/ Encoding/ Collation?
Thanks and Regards,
Shlok Kyal
Attachment | Content-Type | Size |
---|---|---|
test.sh | text/x-sh | 1.4 KB |
From | Date | Subject | |
---|---|---|---|
Next Message | Ranier Vilela | 2024-12-24 10:25:37 | Re: Exists pull-up application with JoinExpr |
Previous Message | torikoshia | 2024-12-24 08:53:06 | RFC: Allow EXPLAIN to Output Page Fault Information |