Re: add timing information to pg_upgrade - Mailing list pgsql-hackers

From Nathan Bossart
Subject Re: add timing information to pg_upgrade
Date
Msg-id 20230802173939.GA1142777@nathanxps13
Whole thread Raw
In response to Re: add timing information to pg_upgrade  (Nathan Bossart <nathandbossart@gmail.com>)
List pgsql-hackers
On Wed, Aug 02, 2023 at 09:09:14AM -0700, Nathan Bossart wrote:
> On Wed, Aug 02, 2023 at 01:02:53PM +0530, Bharath Rupireddy wrote:
>> On Wed, Aug 2, 2023 at 12:45 PM Peter Eisentraut <peter@eisentraut.org> wrote:
>>> I think we should change the output format to be more like initdb, like
>>>
>>>      Doing something ... ok
>>>
>>> without horizontally aligning all the "ok"s.
>> 
>> While this looks simple, we might end up with a lot of diff and
>> changes after removing MESSAGE_WIDTH. There's a significant part of
>> pg_upgrade code that deals with MESSAGE_WIDTH. I don't think it's
>> worth the effort. Therefore, I'd prefer the simplest possible fix -
>> change the message to '"Checking for  \"aclitem\" data type in user
>> tables". It may be an overkill, but we can consider adding
>> Assert(sizeof(message) < MESSAGE_WIDTH) in progress report functions
>> to not encourage new messages to end up in the same formatting issue.
> 
> I don't think it's that difficult.  ІMO the bigger question is whether we
> want to back-patch such a change to v16 at this point.

Here is a work-in-progress patch that seems to get things pretty close to
what Peter is suggesting.

-- 
Nathan Bossart
Amazon Web Services: https://aws.amazon.com

Attachment

pgsql-hackers by date:

Previous
From: Andres Freund
Date:
Subject: Re: SIGQUIT handling, redux
Next
From: Bharath Rupireddy
Date:
Subject: Re: Synchronizing slots from primary to standby