Re: pg_listener in 9.0 - Mailing list pgsql-hackers

From Andrew Dunstan
Subject Re: pg_listener in 9.0
Date
Msg-id 4DE63044.1050507@dunslane.net
Whole thread Raw
In response to Re: pg_listener in 9.0  (Dave Page <dpage@pgadmin.org>)
Responses Re: pg_listener in 9.0
List pgsql-hackers

On 06/01/2011 08:04 AM, Dave Page wrote:
> On Wed, Jun 1, 2011 at 11:27 AM, Heikki Linnakangas
> <heikki.linnakangas@enterprisedb.com>  wrote:
>> On 01.06.2011 13:09, Dave Page wrote:
>>> The pg_listener table was removed in 9.0 in the revamp of
>>> LISTEN/NOTIFY. In pgAdmin we used to perform a number of selects from
>>> the table to get information about Slony clusters - for example, the
>>> PID of the slon process or to check if a process is listening for a
>>> specific notification. This allows the app to indicate to the user if
>>> there is something wrong with their replication cluster.
>>>
>>> I can't find any way to get that information now - any ideas?
>> Hmm, my first thought was that we should add a view to display that
>> information, but that's not possible, because we don't have that information
>> in shared memory. The information on what channels are being listened on is
>> now backend-local.
>>
>> Does the slon process set application_name? You could query pg_stat_activity
>> with that.
> I don't think so (though I might be wrong), but even if it did, it
> wouldn't tell us what cluster it was running against (we figure that
> out by looking at what it's listening for). We also do the same check
> in reverse, to check there is something listening for specific
> notifications.
>

The whole point of the revamp was that pg_listener was a major 
performance bottleneck and needed to go, and without it being gone we 
would not have got notification payloads.

I suspect you're pretty much out of luck.

cheers

andrew


pgsql-hackers by date:

Previous
From: Alexander Korotkov
Date:
Subject: Re: Cube Index Size
Next
From: Teodor Sigaev
Date:
Subject: vacuum and row type