> As for whether it's commonplace, when I was a consultant I had a number > of customers that I had who bemoaned how large updates caused big > replica lag, basically punishing access to records they did care about > in order to properly archive or backfill records they don't care about. > I used the technique a lot, putting the update/delete in a loop, and > often running multiple copies of the same script at times when I/O > contention was low, but if load levels rose it was trivial to just kill > a few of the scripts until things calmed down.
I've also used the technique quite a lot, but only using the PK, didn't know about the ctid trick, so many thanks for documenting it.
tid-scans only became a thing a few versions ago (12?). Prior to that, PK was the only way to go.