Hi !
Indeed :)
Thanks for taking the time to explain !
Benoit.
, February 23, 2017 at 9:23 AM
To: "David G. Johnston"
Cc: Rob Brucks , Tom Lane ,
François Beaulieu , Adrian Klaver ,
"pgsql-general@postgresql.org"
Subject: Re: [GENERAL] NOTIFY command impact
Hi,
I might have missed something.
This should be enough to solve the problem no
Hi,
I might have missed something.
This should be enough to solve the problem no ?
SELECT
CASE
WHEN pg_last_xlog_receive_location() = pg_last_xlog_replay_location() THEN 0
ELSE EXTRACT (EPOCH FROM now() - pg_last_xact_replay_timestamp())::INTEGER
END
AS replication_lag;
Benoit.
On Wed, Feb 22, 2017 at 3:05 PM, Rob Brucks
wrote:
> So, long-term, you don't see any negative impacts to the master cluster?
>
> I just don't want to implement this as a streaming "push" mechanism and
> then have my cluster crash in 12 months because it hit some obscure problem
> with notificati
Thank you Tom.
So, long-term, you don't see any negative impacts to the master cluster?
I just don't want to implement this as a streaming "push" mechanism and then
have my cluster crash in 12 months because it hit some obscure problem with
notifications.
This turned out to be a really simple/
=?utf-8?Q?Fran=C3=A7ois_Beaulieu?= writes:
>> On Feb 21, 2017, at 6:03 PM, Tom Lane wrote:
>> (The reason it goes into the WAL stream is so that you can have listeners
>> on replication slaves, not for recovery purposes.)
> Are we sure that replication slaves can have listeners? When I tried it
> On Feb 21, 2017, at 6:03 PM, Tom Lane wrote:
>
> Rob Brucks writes:
>> If a notify is sent and then PG is immediately shut down, wouldn't PG want
>> to save that message for processing after startup?
>> Or is the message just discarded?
>
> NOTIFY data is not saved across a shutdown or cras
Thanks Tom,
Can you answer my original question too:
If I am issuing NOTIFY commands every 30 seconds (and immediately committing)
and there are no listeners, will that have a negative impact on the cluster?
I'm using the NOTIFY to force streaming replication to update the
"pg_last_xact_replay
> On Feb 21, 2017, at 6:03 PM, Tom Lane wrote:
>
> Rob Brucks writes:
>> If a notify is sent and then PG is immediately shut down, wouldn't PG want
>> to save that message for processing after startup?
>> Or is the message just discarded?
>
> NOTIFY data is not saved across a shutdown or cras
On Tue, Feb 21, 2017 at 3:47 PM, Rob Brucks
wrote:
> If a notify is sent and then PG is immediately shut down, wouldn't PG want
> to save that message for processing after startup?
>
>
>
> Or is the message just discarded?
>
>
>
Adrian gave details but trying to figure out notification semantic
Rob Brucks writes:
> If a notify is sent and then PG is immediately shut down, wouldn't PG want to
> save that message for processing after startup?
> Or is the message just discarded?
NOTIFY data is not saved across a shutdown or crash.
(The reason it goes into the WAL stream is so that you ca
itor the size of :
$PG_DATA/pg_notify/
>
>
>
> --Rob
>
>
>
> *From: *"David G. Johnston"
> *Date: *Tuesday, February 21, 2017 at 3:38 PM
> *To: *Adrian Klaver
> *Cc: *Rob Brucks ,
> "pgsql-general@
"pgsql-general@postgresql.org"
Subject: Re: [GENERAL] NOTIFY command impact
On Tue, Feb 21, 2017 at 3:43 PM, Rob Brucks
mailto:rob.bru...@rackspace.com>> wrote:
But I wonder if that file is only used to store notify commands during
shutdown/startup?
huh?
David J.
On Tue, Feb 21, 2017 at 3:43 PM, Rob Brucks
wrote:
> But I wonder if that file is only used to store notify commands during
> shutdown/startup?
>
huh?
David J.
*Adrian Klaver
> *Cc: *Rob Brucks ,
> "pgsql-general@postgresql.org"
> *Subject: *Re: [GENERAL] NOTIFY command impact
>
>
>
> On Tue, Feb 21, 2017 at 2:17 PM, Adrian Klaver
> mailto:adrian.kla...@aklaver.com>> wrote:
listen does not yet exist.
Seems the thing to do would be to monitor the size of :
$PG_DATA/pg_notify/
--Rob
*From: *"David G. Johnston"
*Date: *Tuesday, February 21, 2017 at 3:38 PM
*To: *Adrian Klaver
*Cc: *Rob Brucks ,
"pgsql-general@postgresql.org"
*Subject: *
David G. Johnston"
Date: Tuesday, February 21, 2017 at 3:38 PM
To: Adrian Klaver
Cc: Rob Brucks , "pgsql-general@postgresql.org"
Subject: Re: [GENERAL] NOTIFY command impact
On Tue, Feb 21, 2017 at 2:17 PM, Adrian Klaver
mailto:adrian.kla...@aklaver.com>> wrote:
On 02/21
In my case the monitoring user will be connecting, issuing the notify, then
immediately disconnecting.
And we don't have any systems using listen/notify.
So I'm hoping there won't be a problem.
That's why I'm asking ☺
--Rob
On 2/21/17, 3:17 PM, "Adrian Klaver" wrote:
On 02/21/2017 01:07
On Tue, Feb 21, 2017 at 2:17 PM, Adrian Klaver
wrote:
> On 02/21/2017 01:07 PM, Rob Brucks wrote:
>
>>
>> Do you see any long-term problems with constantly issuing "NOTIFY"
>> commands every 30 seconds without an associated "LISTEN" command?
>>
>
> Depending on how long 'long term' is:
>
> https:
On 02/21/2017 01:07 PM, Rob Brucks wrote:
Hi All,
I just wanted to check with you guys to make sure that constantly
issuing "NOTIFY" commands without corresponding "LISTEN" commands will
not cause any harm or excessive work for the PostgreSQL cluster. Nothing
that would put my cluster at risk.
20 matches
Mail list logo
