| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | Peter Geoghegan <pg(at)bowt(dot)ie> |
| Cc: | Robert Haas <robertmhaas(at)gmail(dot)com>, PostgreSQL Hackers <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: "Write amplification" is made worse by "getting tired" while inserting into nbtree secondary indexes (Was: Why B-Tree suffix truncation matters) |
| Date: | 2018-07-17 22:10:06 |
| Message-ID: | 1064.1531865406@sss.pgh.pa.us |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Peter Geoghegan <pg(at)bowt(dot)ie> writes:
> I've done plenty of research into the history of this hack. It was
> your work, but it does actually make sense in the context of today's
> nbtree code. It is essential with scankey-wise duplicates, since
> groveling through hundreds or even thousands of pages full of
> duplicates to find free space (and avoid a page split) is going to
> have a very serious downside for latency.
Well, the actual problem was O(N^2) behavior:
https://www.postgresql.org/message-id/2378.967216388%40sss.pgh.pa.us
I certainly have no objection to improving matters, but let's be sure
we don't re-introduce any two-decade-old problems.
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | R, Siva | 2018-07-17 22:40:12 | Bug in gin insert redo code path during re-compression of empty gin data leaf pages |
| Previous Message | Simon Riggs | 2018-07-17 22:00:18 | Re: Allowing multiple DDL commands to run simultaneously |