Re: (replication) Detecting if server a slave, or a master in recovery
От | Guillaume Lelarge |
---|---|
Тема | Re: (replication) Detecting if server a slave, or a master in recovery |
Дата | |
Msg-id | 1315815982.2708.7.camel@localhost.localdomain обсуждение исходный текст |
Ответ на | Re: (replication) Detecting if server a slave, or a master in recovery (Toby Corkindale <toby.corkindale@strategicdata.com.au>) |
Список | pgsql-general |
On Mon, 2011-09-12 at 17:09 +1000, Toby Corkindale wrote: > On 12/09/11 16:54, Toby Corkindale wrote: > > Hi, > > What is the correct way to tell what mode a replicated Pg server is > > currently in, via a database connection? > > > > I can take a guess if it's the master or a slave by using the > > pg_current_xlog_location() and pg_last_xlog_replay_location() functions. > > > > However it occurs to me that a master server, which has been roughly > > rebooted and is replaying its WAL files will probably give the same > > response as a slave, when I call the current_xlog_location(). > > > > > > What is the best method for determining whether you're connected to the > > master or slave database? > > > It's disappointing that I can't query standby_mode in psql. > ie. > SHOW standby_mode; > I agree. Actually, you can't get the value of any parameter set in recovery.conf. But Fuji Masao seems to work on it: he posted a patch to unite recovery.conf and postgresql.conf (see "unite recovery.conf and postgresql.conf" thread on pgsql-hackers). And I guess it'll help us querying these parameters' values. -- Guillaume http://blog.guillaume.lelarge.info http://www.dalibo.com
В списке pgsql-general по дате отправления: