On Sat, Jan 14, 2012 at 5:42 AM, Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
> Other than that, it seems like we might be converging on a workable
> solution: if hot_standby_feedback=off, disable index-only scans for
> snapshots taken during recovery; if hot_standby_feedback=on, generate
> recovery conflicts when a snapshot's xmin precedes the youngest xmin
> on a page marked all-visible, but allow the use of index-only scans,
> and allow sequential scans to trust PD_ALL_VISIBLE. Off the top of my
> head, I don't see a hole in that logic...
Please use the logic described earlier. Using the existing mechanisms
and infrastructure is all we need here.
We want latestRemovedXid, not a new version of it via xmins. Keep
index-only scans enabled all the time and let conflicts be generated.
There are already mechanisms in place to handle conflicts, so avoiding
them another way is not required or desired.
Simon Riggs http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
In response to
pgsql-hackers by date
|Next:||From: Jaime Casanova||Date: 2012-01-14 09:41:57|
|Subject: Re: Measuring relation free space|
|Previous:||From: Simon Riggs||Date: 2012-01-14 08:08:29|
|Subject: Re: Disabled features on Hot Standby|