Re: Patch (2): Implement failover on libpq connect level. - Mailing list pgsql-hackers

From Victor Wagner
Subject Re: Patch (2): Implement failover on libpq connect level.
Date
Msg-id 20151023145233.2a110244@fafnir
Whole thread Raw
In response to Re: Patch (2): Implement failover on libpq connect level.  (Thom Brown <thom@linux.com>)
Responses Re: Patch (2): Implement failover on libpq connect level.
List pgsql-hackers
On Thu, 22 Oct 2015 14:33:11 +0100
Thom Brown <thom@linux.com> wrote:

> On 21 October 2015 at 10:07, Victor Wagner <vitus@wagner.pp.ru> wrote:
> > On 2015.10.14 at 13:41:51 +0300, Victor Wagner wrote:
> >
> >>
> >> Attached patch which implements client library failover and
> >> loadbalancing as was described in the proposal
> >> <20150818041850.GA5092@wagner.pp.ru>.
> >>
> >
> > I'm sending imporoved verison of patch. As Olexander Shulgin noted,
> > previous version of patch lacks support for service files.
> >
> > Now support for service files is implemented and multiple host
> > statements in the service file are allowed.
> 
> This is causing breakage:
> 
> $ pg_basebackup -v -x -D standby1 -h localhost -p 5532 -U rep_user
> row number 0 is out of range 0..-1

It seems that pg_basebackup should always specify readonly attribute
for the connection. 

Your data directory is named standby1, so I suspect
that you are trying to make backup from read-only standby instance of
the base. 

Can you check if problem persists in your setup with command

pg_basebackup -v -x -D standby1 \ -d "host=localhost port=5532 user=rep_user readonly=1"


> Segmentation fault

Of course, this shouldn't happen even if libpq is severely misused.
And it is almost evident what should be done to fix it.
But making sure that my patch doesn't interfere with operation of
pg_basebackup is a bit more complicated.





pgsql-hackers by date:

Previous
From: Amit Kapila
Date:
Subject: Re: Parallel Seq Scan
Next
From: Michael Paquier
Date:
Subject: Re: [patch] extensions_path GUC