| From: | Peter Geoghegan <pg(at)bowt(dot)ie> |
|---|---|
| To: | Simon Riggs <simon(dot)riggs(at)enterprisedb(dot)com> |
| Cc: | Robert Haas <robertmhaas(at)gmail(dot)com>, Matthias van de Meent <boekewurm+postgres(at)gmail(dot)com>, John Naylor <john(dot)naylor(at)enterprisedb(dot)com>, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, Mark Dilger <mark(dot)dilger(at)enterprisedb(dot)com>, PostgreSQL Hackers <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: Lowering the ever-growing heap->pd_lower |
| Date: | 2021-08-05 16:14:59 |
| Message-ID: | CAH2-WznpQ=KkSfd6GmaWEezS=2h40szF3OE1RsV-aLBCATGnNw@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Thu, Aug 5, 2021 at 6:28 AM Simon Riggs <simon(dot)riggs(at)enterprisedb(dot)com> wrote:
> Hmm, there is no information in WAL to describe the line pointers
> being truncated by PageTruncateLinePointerArray(). We just truncate
> every time we see a XLOG_HEAP2_VACUUM record and presume it does the
> same thing as the original change.
>
> If that is safe, then we don't need to put the truncation on a WAL
> record at all, we just truncate after every XLOG_HEAP2_PRUNE record.
I agree that that's how we'd do it. This approach is no different to
assuming that PageRepairFragmentation() reliably produces a final
defragmented page image deterministically when called after we prune.
These days we automatically verify assumptions like this via
wal_consistency_checking. It would absolutely be able to catch any
bugs in PageTruncateLinePointerArray(), since the truncate code path
has plenty of coverage within the regression tests.
--
Peter Geoghegan
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Dipesh Pandit | 2021-08-05 16:23:30 | Re: .ready and .done files considered harmful |
| Previous Message | Tom Lane | 2021-08-05 16:04:39 | Re: Numeric x^y for negative x |