Re: bug: copy progress reporting of backends which run multiple COPYs

Matthias van de Meent <boekewurm+postgres@gmail.com>

From: Matthias van de Meent <boekewurm+postgres@gmail.com>

To: Justin Pryzby <pryzby@telsasoft.com>

Cc: Tomas Vondra <tomas.vondra@enterprisedb.com>, Josef Šimánek <josef.simanek@gmail.com>, Michael Paquier <michael@paquier.xyz>, pgsql-hackers@lists.postgresql.org

Date: 2023-01-21T00:51:28Z

Lists: pgsql-hackers

Attachments

v1-0001-Only-report-progress-if-we-re-actually-in-a-parse.patch (application/octet-stream) patch v1-0001

On Thu, 19 Jan 2023 at 06:47, Justin Pryzby <pryzby@telsasoft.com> wrote:
>
> pg_stat_progress_copy was added in v14 (8a4f618e7, 9d2d45700).
>
> But if a command JOINs file_fdw tables, the progress report gets bungled
> up.  This will warn/assert during file_fdw tests.

I don't know what to do with that other than disabling COPY progress
reporting for file_fdw, i.e. calls to BeginCopyFrom that don't supply
a pstate. This is probably the best option, because a table backed by
file_fdw would also interfere with COPY TO's progress reporting.

Attached a patch that solves this specific issue in a
binary-compatible way. I'm not super happy about relying on behavior
of callers of BeginCopyFrom (assuming that users that run copy
concurrently will not provide a ParseState* to BeginCopyFrom), but it
is what it is.

Kind regards,

Matthias van de Meent