Horton

Re: improve performance of pg_dump with many sequences

Euler Taveira <euler@eulerto.com>

View thread View on postgr.es
From: "Euler Taveira" <euler@eulerto.com>
To: "Nathan Bossart" <nathandbossart@gmail.com>
Cc: pgsql-hackers <pgsql-hackers@postgresql.org>
Date: 2024-07-11T02:52:33Z
Lists: pgsql-hackers 
On Wed, Jul 10, 2024, at 7:05 PM, Nathan Bossart wrote:
> I'm not following why that would be a better approach.  strncpy() will add
> a NUL to the end of the string unless it doesn't fit in the buffer, in
> which case we'll add our own via "seqtype[sizeof(seqtype) - 1] = '\0'".
> Furthermore, the compiler can determine the position where the NUL should
> be placed, whereas placing it at the end of the copied string requires a
> runtime strlen().

Nevermind, you are copying the whole buffer (n = sizeof(seqtype)).

> Unfortunately, I think we have to keep this workaround since older minor
> releases of PostgreSQL don't have the fix.

Hmm. Right.

> What pg_dump command did you test here?  Did you dump the sequence data, or
> was this --schema-only?

time pg_dump -f - -s -d postgres


--
Euler Taveira
EDB   https://www.enterprisedb.com/

Commits

  1. pg_dump: Fix gathering of sequence information.

  2. Improve performance of dumpSequence().

  3. Improve performance of dumpSequenceData().

  4. Introduce pg_sequence_read_tuple().

  5. Parse sequence type and integer metadata in dumpSequence().