From: | "Drouvot, Bertrand" <bertranddrouvot(dot)pg(at)gmail(dot)com> |
---|---|
To: | Amit Kapila <amit(dot)kapila16(at)gmail(dot)com>, Nathan Bossart <nathandbossart(at)gmail(dot)com> |
Cc: | PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org> |
Subject: | Re: Change xl_hash_vacuum_one_page.ntuples from int to uint16 |
Date: | 2023-02-17 14:13:17 |
Message-ID: | 3902292e-cf56-17d5-4b61-d883c82c9f9a@gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Hi,
On 2/16/23 1:26 PM, Drouvot, Bertrand wrote:
> Hi,
>
> On 2/16/23 12:00 PM, Amit Kapila wrote:
>
>> BTW, feel free to create the second patch
>> (to align the types for variables/arguments) as that would be really
>> helpful after we commit this one.
Please find attached a patch proposal to do so.
It looks like a Pandora's box as it produces
those cascading changes:
_hash_vacuum_one_page
index_compute_xid_horizon_for_tuples
gistprunepage
PageIndexMultiDelete
gistXLogDelete
PageIndexMultiDelete
spgRedoVacuumRedirect
vacuumRedirectAndPlaceholder
spgPageIndexMultiDelete
moveLeafs
doPickSplit
_bt_delitems_vacuum
btvacuumpage
_bt_delitems_delete
_bt_delitems_delete_check
hash_xlog_move_page_contents
gistvacuumpage
gistXLogUpdate
gistplacetopage
hashbucketcleanup
Which makes me:
- wonder it is not too intrusive (we could reduce the scope and keep the
PageIndexMultiDelete()'s nitems argument as an int for example).
- worry if there is no side effects (like the one I'm mentioning as a comment
in PageIndexMultiDelete()) even if it passes the CI tests.
- wonder if we should not change MaxIndexTuplesPerPage from int to uint16 too (given
the fact that the maximum block size is 32 KB.
I'm sharing this version but I still need to think about it and
I'm curious about your thoughts too.
Regards,
--
Bertrand Drouvot
PostgreSQL Contributors Team
RDS Open Source Databases
Amazon Web Services: https://aws.amazon.com
Attachment | Content-Type | Size |
---|---|---|
v1-0001-Change-ndeletable-in-_hash_vacuum_one_page-from-i.patch | text/plain | 15.7 KB |
From | Date | Subject | |
---|---|---|---|
Next Message | Peter Eisentraut | 2023-02-17 14:13:32 | Re: DDL result is lost by CREATE DATABASE with WAL_LOG strategy |
Previous Message | Jonathan S. Katz | 2023-02-17 14:06:09 | Re: The output sql generated by pg_dump for a create function refers to a modified table name |