| From: | "Hayato Kuroda (Fujitsu)" <kuroda(dot)hayato(at)fujitsu(dot)com> |
|---|---|
| To: | "maxim(dot)boguk(at)gmail(dot)com" <maxim(dot)boguk(at)gmail(dot)com> |
| Cc: | "pgsql-bugs(at)lists(dot)postgresql(dot)org" <pgsql-bugs(at)lists(dot)postgresql(dot)org> |
| Subject: | RE: BUG #18644: ALTER PUBLICATION ... SET (publish_via_partition_root) wrong/undocumented behavior. |
| Date: | 2024-10-04 04:30:13 |
| Message-ID: | TYAPR01MB56925E13BC85C58C203F8902F5722@TYAPR01MB5692.jpnprd01.prod.outlook.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-bugs |
Dear Maxim,
Thanks for reporting the issue. Before discussing more about it, I want to confirm
the exact workload you did. Is below sequence same as you expected?
0. There is a table which have leaf tables on publisher/subscriber.
1. create a PUBLICATION and include a root table
2. create a SUBSCRIPTION to subscribe above
3. insert some data on pub and wait until the replication is done.
4. do ALTER PUBLICAITON WITH (publish_via_partition_root = true)
5. insert tuples on the publisher
-> tuples won't be replicated until ALTER SUBSCRIPTION REFRESH PUBLICATION is done.
If the copy_data is true for REFRESH, all tuples inserted at 3. will be copied again.
Otherwise, data between ALTER PUB. - ALTER SUB. will be lost.
Attached .pl file (renamed to .txt due to the security reason) can reproduce and
test the situation via our test framework.
Best regards,
Hayato Kuroda
FUJITSU LIMITED
| Attachment | Content-Type | Size |
|---|---|---|
| 034_alter_pub.txt | text/plain | 2.8 KB |
| From | Date | Subject | |
|---|---|---|---|
| Next Message | neral85 | 2024-10-04 06:15:43 | AW: BUG #18615: installer cannot be executed as "nt-autorität\system" |
| Previous Message | Noah Misch | 2024-10-04 02:33:37 | Re: BUG #18637: CREATE INDEX won't look up operator classes in search_path if PARTITION BY is specified |