[Pgpool-hackers] Enhancing show pool_nodes
Guillaume Lelarge
guillaume at lelarge.info
Fri Oct 7 07:00:43 UTC 2011
On Fri, 2011-10-07 at 15:30 +0900, Tatsuo Ishii wrote:
> > On Thu, 2011-10-06 at 10:10 +0900, Tatsuo Ishii wrote:
> >> Currently show pool_nodes SQL command shows:
> >>
> >> id | hostname | port | status | lb_weight
> >>
> >> I would like to add info whether the node is primary, master or
> >> standby because there's no way to know from exisiting pgpool-II UI.
> >> The new column name would be "class" (plese say no, if you have better
> >> idea about naming). The value will as follows:
> >>
> >> Raw mode:
> >> "master" or ""
> >>
> >> Native replication mode:
> >> "master" or "slave"
> >>
> >> Streaming replication mode:
> >> "primary" or "standby"
> >>
> >> Master slave mode with slony:
> >> "master" or "slave"
> >>
> >> Comments or opinions?
> >
> > I guess it would be better to have the same vocabulary. So I would vote
> > "master"/"slave" for each replication mode, and "master" for raw mode.
>
> Hum. I'd prefer to use "primary" and "standby" to keep consistency
> with PostgreSQL's terminology. We might use even another word for 9.1,
> for example "synchronous standby". So just only using "master/slave"
> would be more confusing and less informative to user, I think.
The same will happen to Slony users: they use "master" and "slave", or
"provider"/"subscriber", but never "primary"/"standby". PostgreSQL users
use both (but never "provider"/"subscriber").
So, "master"/"slave" is more generic terms than "primary"/"standby".
--
Guillaume
http://blog.guillaume.lelarge.info
http://www.dalibo.com
More information about the Pgpool-hackers
mailing list