Re: Allow "snapshot too old" error, to prevent bloat

<p dir="ltr"><br /> On Feb 17, 2015 12:26 AM, "Andres Freund" <<a
href="mailto:andres@2ndquadrant.com">andres@2ndquadrant.com</a>>wrote:<br /> ><br /> > On 2015-02-16 16:35:46
-0500,Bruce Momjian wrote:<br /> > > It seems we already have a mechanism in place that allows tuning of<br />
>> query cancel on standbys vs. preventing standby queries from seeing old<br /> > > data, specifically<br
/>> > max_standby_streaming_delay/max_standby_archive_delay.  We obsessed<br /> > > about how users were
goingto react to these odd variables, but there<br /> > > has been little negative feedback.<br /> ><br />
>FWIW, I think that's a somewhat skewed perception. I think it was right to<br /> > introduce those, because we
didn'treally have any alternatives.<br /> ><br /> > But max_standby_streaming_delay, max_standby_archive_delay
and<br/> > hot_standby_feedback are among the most frequent triggers for questions<br /> > and complaints that
I/wesee.<br /> ><p dir="ltr">Agreed.<p dir="ltr">And a really bad one used to be vacuum_defer_cleanup_age, because
ofconfusing units amongst other things. Which in terms seems fairly close to Kevins suggestions, unfortunately. <br
/><pdir="ltr">/Magnus <br />