Re: Enable data checksums by default

Peter Eisentraut <peter@eisentraut.org>

From: Peter Eisentraut <peter@eisentraut.org>

To: Andrew Dunstan <andrew@dunslane.net>

Cc: Nathan Bossart <nathandbossart@gmail.com>, Greg Sabino Mullane <htamfids@gmail.com>, Bruce Momjian <bruce@momjian.us>, Michael Paquier <michael@paquier.xyz>, Michael Banck <mbanck@gmx.net>, pgsql-hackers <pgsql-hackers@postgresql.org>

Date: 2024-10-16T07:49:48Z

Lists: pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

Avoid BufferGetLSNAtomic() calls during nbtree scans.
- e6eed40e4441 18.0 cited
doc PG 18 relnotes: Add incompatibility note about checksums now default
- 48814415d5aa 18.0 landed
Fix pg_upgrade's cross-version tests when old < 18
- 38c18710b37a 18.0 landed
initdb: Change default to using data checksums.
- 04bec894a04c 18.0 landed
Allow TAP tests to force checksums off when calling init()
- e7d0cf42b1ac 18.0 landed
initdb: Add new option "--no-data-checksums"
- 983a588e0b86 18.0 landed
Tweak docs to reduce possible impact of data checksums
- efd72a3d422b 18.0 landed

On 16.10.24 08:54, Peter Eisentraut wrote:
> On 14.10.24 11:28, Peter Eisentraut wrote:
>> On 03.10.24 23:13, Nathan Bossart wrote:
>>> On Tue, Oct 01, 2024 at 11:15:02AM -0400, Peter Eisentraut wrote:
>>>> I have committed 0001 (the new option) and 0004 (the docs tweak).  I 
>>>> think
>>>> there is consensus for the rest, too, but I'll leave it for a few 
>>>> more days
>>>> to think about.  I guess the test failure has to be addressed.
>>>
>>> Here is a rebased patch with the test fix (for cfbot).  I have made no
>>> other changes.
>>
>> I have committed the test changes (patch 0002).  (I renamed the option 
>> to no_data_checksums to keep the wording consistent with the initdb 
>> option.)
>>
>> Right now, with checksums off by default, this doesn't do much, but 
>> you can test this like
>>
>>      PG_TEST_INITDB_EXTRA_OPTS=--data-checksums meson test ...
>>
>> and everything will pass.  To make that work, I had to adjust the 
>> order of how the initdb options are assembled in Cluster.pm a bit.
>>
>> I will work on the patch that flips the default next.
> 
> The patch that flips the default has been committed.
> 
> I also started a PG18 open items page and made a note that we follow up 
> on the upgrade experience, as was discussed in this thread.

Ah yes, and the upgrade tests on the buildfarm don't like this.  What 
shall we do about this?  Maybe adjust the buildfarm scripts to use 
--no-data-checksums?