On Wed, Mar 29, 2017 at 04:58:40PM -0300, Alvaro Herrera wrote:
> Robert Haas wrote:
> > On Wed, Mar 29, 2017 at 2:10 PM, Peter Eisentraut
> > <peter.eisentraut@2ndquadrant.com> wrote:
> > > How specifically would we do that? And what user would choose the
> > > behavior "start this background worker but don't worry if it doesn't work"?
> >
> > Well, if the background worker is auto-prewarm, you'd probably rather
> > have the database start rather than get unhappy about auto-prewarm
> > failing. If the background worker is your logical replication
> > launcher it's a bit more serious, but if you have no subscriptions or
> > they're not that critical, maybe you don't care. If the background
> > worker is in charge of telling your failover solution that this node
> > is up, then starting without it is entirely pointless.
> >
> > I would be inclined to leave this alone for now and revisit it for a
> > future release. I don't feel confident that we really know what the
> > right thing to do is here.
>
> I think the common case is for modules to be critical: you may not care
> about it for auto-prewarm, but that seems like a special case. I would
> put it the other way around: having the load fail is a serious problem
> unless specifically configured not to be. I'd do as Peter suggests, and
> perhaps allow the current behavior optionally. In hindsight, the
> current behavior seems like a mistake.
Agreed. There are times when starting up degraded beats failing to start,
particularly when the failing component has complicated dependencies. In this
case, as detailed upthread, this can only fail in response to basic
misconfiguration. It's not the kind of thing that will spontaneously fail
after a minor upgrade, for example.
[Action required within three days. This is a generic notification.]
The above-described topic is currently a PostgreSQL 10 open item. Peter,
since you committed the patch believed to have created it, you own this open
item. If some other commit is more relevant or if this does not belong as a
v10 open item, please let us know. Otherwise, please observe the policy on
open item ownership[1] and send a status update within three calendar days of
this message. Include a date for your subsequent status update. Testers may
discover new open items at any time, and I want to plan to get them all fixed
well in advance of shipping v10. Consequently, I will appreciate your efforts
toward speedy resolution. Thanks.
[1] https://www.postgresql.org/message-id/20170404140717.GA2675809%40tornado.leadboat.com
