pg_stat_replication vs StandbyReplyMessage

Started by Magnus Haganderover 14 years ago10 messages
#1Magnus Hagander
magnus@hagander.net

The pg_stat_replication view exposes all the fields in
StandbyReplyMessage *except* for the timestamp when the message was
generated. On an active system this is not all that interesting, but
on a mostly idle system that allows the monitoring to react faster
than the timeout that actually kicks the other end off - and could be
useful in manual debugging scenarios. Any particular reason why this
was not exposed as it's own column?

--
 Magnus Hagander
 Me: http://www.hagander.net/
 Work: http://www.redpill-linpro.com/

#2Robert Haas
robertmhaas@gmail.com
In reply to: Magnus Hagander (#1)
Re: pg_stat_replication vs StandbyReplyMessage

On Mon, Aug 15, 2011 at 7:03 AM, Magnus Hagander <magnus@hagander.net> wrote:

The pg_stat_replication view exposes all the fields in
StandbyReplyMessage *except* for the timestamp when the message was
generated. On an active system this is not all that interesting, but
on a mostly idle system that allows the monitoring to react faster
than the timeout that actually kicks the other end off - and could be
useful in manual debugging scenarios. Any particular reason why this
was not exposed as it's own column?

I wondered the same thing. Sounds like a good idea.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

#3Magnus Hagander
magnus@hagander.net
In reply to: Robert Haas (#2)
Re: pg_stat_replication vs StandbyReplyMessage

On Mon, Aug 15, 2011 at 13:50, Robert Haas <robertmhaas@gmail.com> wrote:

On Mon, Aug 15, 2011 at 7:03 AM, Magnus Hagander <magnus@hagander.net> wrote:

The pg_stat_replication view exposes all the fields in
StandbyReplyMessage *except* for the timestamp when the message was
generated. On an active system this is not all that interesting, but
on a mostly idle system that allows the monitoring to react faster
than the timeout that actually kicks the other end off - and could be
useful in manual debugging scenarios. Any particular reason why this
was not exposed as it's own column?

I wondered the same thing.  Sounds like a good idea.

I can go do that. Care to argue^Wbikeshed for a specific name?

--
 Magnus Hagander
 Me: http://www.hagander.net/
 Work: http://www.redpill-linpro.com/

#4Simon Riggs
simon@2ndQuadrant.com
In reply to: Magnus Hagander (#3)
Re: pg_stat_replication vs StandbyReplyMessage

On Tue, Aug 16, 2011 at 10:51 AM, Magnus Hagander <magnus@hagander.net> wrote:

On Mon, Aug 15, 2011 at 13:50, Robert Haas <robertmhaas@gmail.com> wrote:

On Mon, Aug 15, 2011 at 7:03 AM, Magnus Hagander <magnus@hagander.net> wrote:

The pg_stat_replication view exposes all the fields in
StandbyReplyMessage *except* for the timestamp when the message was
generated. On an active system this is not all that interesting, but
on a mostly idle system that allows the monitoring to react faster
than the timeout that actually kicks the other end off - and could be
useful in manual debugging scenarios. Any particular reason why this
was not exposed as it's own column?

I wondered the same thing.  Sounds like a good idea.

I can go do that. Care to argue^Wbikeshed for a specific name?

reply_timestamp

--
 Simon Riggs                   http://www.2ndQuadrant.com/
 PostgreSQL Development, 24x7 Support, Training & Services

#5Robert Haas
robertmhaas@gmail.com
In reply to: Simon Riggs (#4)
Re: pg_stat_replication vs StandbyReplyMessage

On Tue, Aug 16, 2011 at 8:54 AM, Simon Riggs <simon@2ndquadrant.com> wrote:

On Tue, Aug 16, 2011 at 10:51 AM, Magnus Hagander <magnus@hagander.net> wrote:

On Mon, Aug 15, 2011 at 13:50, Robert Haas <robertmhaas@gmail.com> wrote:

On Mon, Aug 15, 2011 at 7:03 AM, Magnus Hagander <magnus@hagander.net> wrote:

The pg_stat_replication view exposes all the fields in
StandbyReplyMessage *except* for the timestamp when the message was
generated. On an active system this is not all that interesting, but
on a mostly idle system that allows the monitoring to react faster
than the timeout that actually kicks the other end off - and could be
useful in manual debugging scenarios. Any particular reason why this
was not exposed as it's own column?

I wondered the same thing.  Sounds like a good idea.

I can go do that. Care to argue^Wbikeshed for a specific name?

reply_timestamp

Works for me. I'd suggest that we rename it that way in
StandbyReplyMessage, so that the name in the struct and the name in
the system view match.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

#6Simon Riggs
simon@2ndQuadrant.com
In reply to: Robert Haas (#5)
Re: pg_stat_replication vs StandbyReplyMessage

On Tue, Aug 16, 2011 at 2:25 PM, Robert Haas <robertmhaas@gmail.com> wrote:

On Tue, Aug 16, 2011 at 8:54 AM, Simon Riggs <simon@2ndquadrant.com> wrote:

On Tue, Aug 16, 2011 at 10:51 AM, Magnus Hagander <magnus@hagander.net> wrote:

On Mon, Aug 15, 2011 at 13:50, Robert Haas <robertmhaas@gmail.com> wrote:

On Mon, Aug 15, 2011 at 7:03 AM, Magnus Hagander <magnus@hagander.net> wrote:

The pg_stat_replication view exposes all the fields in
StandbyReplyMessage *except* for the timestamp when the message was
generated. On an active system this is not all that interesting, but
on a mostly idle system that allows the monitoring to react faster
than the timeout that actually kicks the other end off - and could be
useful in manual debugging scenarios. Any particular reason why this
was not exposed as it's own column?

I wondered the same thing.  Sounds like a good idea.

I can go do that. Care to argue^Wbikeshed for a specific name?

reply_timestamp

Works for me.

I'd suggest that we rename it that way in
StandbyReplyMessage, so that the name in the struct and the name in
the system view match.

-1

The field is named same as equivalent field in other messages.

The field on the view is a summary of all previous messages, which is
a different thing. Perhaps we should call it last_reply_timestamp to
make that clearer, though long titles are annoying.

--
 Simon Riggs                   http://www.2ndQuadrant.com/
 PostgreSQL Development, 24x7 Support, Training & Services

#7Magnus Hagander
magnus@hagander.net
In reply to: Simon Riggs (#6)
Re: pg_stat_replication vs StandbyReplyMessage

On Tue, Aug 16, 2011 at 16:00, Simon Riggs <simon@2ndquadrant.com> wrote:

On Tue, Aug 16, 2011 at 2:25 PM, Robert Haas <robertmhaas@gmail.com> wrote:

On Tue, Aug 16, 2011 at 8:54 AM, Simon Riggs <simon@2ndquadrant.com> wrote:

On Tue, Aug 16, 2011 at 10:51 AM, Magnus Hagander <magnus@hagander.net> wrote:

On Mon, Aug 15, 2011 at 13:50, Robert Haas <robertmhaas@gmail.com> wrote:

On Mon, Aug 15, 2011 at 7:03 AM, Magnus Hagander <magnus@hagander.net> wrote:

The pg_stat_replication view exposes all the fields in
StandbyReplyMessage *except* for the timestamp when the message was
generated. On an active system this is not all that interesting, but
on a mostly idle system that allows the monitoring to react faster
than the timeout that actually kicks the other end off - and could be
useful in manual debugging scenarios. Any particular reason why this
was not exposed as it's own column?

I wondered the same thing.  Sounds like a good idea.

I can go do that. Care to argue^Wbikeshed for a specific name?

reply_timestamp

Works for me.

I'd suggest that we rename it that way in
StandbyReplyMessage, so that the name in the struct and the name in
the system view match.

-1

The field is named same as equivalent field in other messages.

The field on the view is a summary of all previous messages, which is
a different thing. Perhaps we should call it last_reply_timestamp to
make that clearer, though long titles are annoying.

We don't say last_replay_location either, we just say replay_location.
Adding the last_ part is just annoying.

--
 Magnus Hagander
 Me: http://www.hagander.net/
 Work: http://www.redpill-linpro.com/

#8Bruce Momjian
bruce@momjian.us
In reply to: Magnus Hagander (#1)
Re: pg_stat_replication vs StandbyReplyMessage

On Mon, Aug 15, 2011 at 01:03:35PM +0200, Magnus Hagander wrote:

The pg_stat_replication view exposes all the fields in
StandbyReplyMessage *except* for the timestamp when the message was
generated. On an active system this is not all that interesting, but
on a mostly idle system that allows the monitoring to react faster
than the timeout that actually kicks the other end off - and could be
useful in manual debugging scenarios. Any particular reason why this
was not exposed as it's own column?

Did this ever get done? I don't think so, though everyone wanted it.

--
Bruce Momjian <bruce@momjian.us> http://momjian.us
EnterpriseDB http://enterprisedb.com

+ It's impossible for everything to be true. +

#9Magnus Hagander
magnus@hagander.net
In reply to: Bruce Momjian (#8)
Re: pg_stat_replication vs StandbyReplyMessage

On Wed, Aug 15, 2012 at 5:39 PM, Bruce Momjian <bruce@momjian.us> wrote:

On Mon, Aug 15, 2011 at 01:03:35PM +0200, Magnus Hagander wrote:

The pg_stat_replication view exposes all the fields in
StandbyReplyMessage *except* for the timestamp when the message was
generated. On an active system this is not all that interesting, but
on a mostly idle system that allows the monitoring to react faster
than the timeout that actually kicks the other end off - and could be
useful in manual debugging scenarios. Any particular reason why this
was not exposed as it's own column?

Did this ever get done? I don't think so, though everyone wanted it.

Nope, it wasn't done. Should probably do that for 9.3 (since adding a
field to pg_stat_replication will cause initdb, so we can't really do
it for 9.2 unless it was really critical - and it's not).

--
Magnus Hagander
Me: http://www.hagander.net/
Work: http://www.redpill-linpro.com/

#10Bruce Momjian
bruce@momjian.us
In reply to: Magnus Hagander (#9)
Re: pg_stat_replication vs StandbyReplyMessage

On Thu, Aug 23, 2012 at 12:40:33PM +0200, Magnus Hagander wrote:

On Wed, Aug 15, 2012 at 5:39 PM, Bruce Momjian <bruce@momjian.us> wrote:

On Mon, Aug 15, 2011 at 01:03:35PM +0200, Magnus Hagander wrote:

The pg_stat_replication view exposes all the fields in
StandbyReplyMessage *except* for the timestamp when the message was
generated. On an active system this is not all that interesting, but
on a mostly idle system that allows the monitoring to react faster
than the timeout that actually kicks the other end off - and could be
useful in manual debugging scenarios. Any particular reason why this
was not exposed as it's own column?

Did this ever get done? I don't think so, though everyone wanted it.

Nope, it wasn't done. Should probably do that for 9.3 (since adding a
field to pg_stat_replication will cause initdb, so we can't really do
it for 9.2 unless it was really critical - and it's not).

OK, TODO added:

Add entry creation timestamp column to pg_stat_replication

http://archives.postgresql.org/pgsql-hackers/2011-08/msg00694.php

--
Bruce Momjian <bruce@momjian.us> http://momjian.us
EnterpriseDB http://enterprisedb.com

+ It's impossible for everything to be true. +