hi.
within ATRewriteTable we have:
pgstat_progress_update_param(PROGRESS_CLUSTER_TOTAL_HEAP_BLKS,
RelationGetNumberOfBlocks(oldrel));
pgstat_progress_update_param(PROGRESS_CLUSTER_TOTAL_HEAP_BLKS,
heapScan->rs_nblocks);
PROGRESS_CLUSTER_TOTAL_HEAP_BLKS
value is fixed, we only need to call pgstat_progress_update_param once here?
another patch [1] is expected to refactor pg_stat_progress_cluster a lot, so I'm
unsure whether it's a good idea to put CLUSTER, VACUUM FULL, or ALTER TABLE into
pg_stat_progress_cluster.
alternatively, we could introduce a separate progress report specifically for
ALTER TABLE, allowing us to distinguish between table rewrite and table
scan.
[1] https://commitfest.postgresql.org/patch/5117