[pgpool-general: 3090] Re: timeout Watchdog

Yugo Nagata nagata at sraoss.co.jp
Thu Jul 31 21:56:46 JST 2014


Hi,

I'm so sorry for delayed reply.
I confirmed that watchdog lifecheck didn't start. If this don't start after
waiting enough, configure of hearbeat might be somowhat wrong.

On Fri, 25 Apr 2014 16:37:18 -0300
Thomaz Luiz Santos <thomaz.santos at gmail.com> wrote:

> I see in the standby log, is a problem in the failover script?
> 
> 
> On Wed, Apr 23, 2014 at 10:21 PM, Yugo Nagata <nagata at sraoss.co.jp> wrote:
> 
> > On Wed, 23 Apr 2014 17:25:10 -0300
> > Thomaz Luiz Santos <thomaz.santos at gmail.com> wrote:
> >
> > > yes, I disconnected the lan on "active".
> > >
> > > in my LOG, dont have watchdog lifecheck started :-(
> > >
> > > on "active" / Master
> > > root at master:~# cat /var/log/syslog | grep watchdog
> > > Apr 23 13:12:11 master pgpool: 2014-04-23 13:12:11 LOG:   pid 2697:
> > > wd_init: start watchdog
> > >
> > > on "standby" / slave
> > > root at slave:~# cat /var/log/syslog | grep watchdog
> > > Apr 23 13:13:08 slave pgpool: 2014-04-23 13:13:08 LOG:   pid 2460:
> > wd_init:
> > > start watchdog
> > >
> > > I use the tutorial :-D
> > >
> > > How I can start the lifecheck ?
> >
> > It cant take about 50 seconds (wd_interval * 10) from starting of
> > standby pgpool. Please wait till the message appears.
> >
> > If you can't see after waiting enough, there would be problem with
> > configuration or pgpool itself. So, for analysis could you please
> > show me the all log messages?
> >
> > >
> > >
> > > On Tue, Apr 22, 2014 at 10:41 PM, Yugo Nagata <nagata at sraoss.co.jp>
> > wrote:
> > >
> > > > Hi,
> > > >
> > > > On Tue, 22 Apr 2014 16:42:42 -0300
> > > > Thomaz Luiz Santos <thomaz.santos at gmail.com> wrote:
> > > >
> > > > > I changed the parameters,
> > > > > and disconnected the network cable
> > > > > but the slave not change to master
> > > >
> > > > You mean that master is pgpool which has VIP (this is called "active")
> > > > and slave is the other pgpool (this is called "standby"),
> > > > and that you disconnected the LAN cable of the active pgpool,
> > > > is this right?
> > > >
> > > > >
> > > > > on pgpool.conf:
> > > > > wd_interval = 5
> > > > > wd_life_point = 5
> > > >
> > > > Hmm, the standby pgpool should become active in about 30 seconds with
> > > > this parameters.... Maybe, watchdog lifecheck might not have started
> > yet.
> > > > Confirm that the following message appears in log messages before
> > > > disconneting the cable.
> > > >
> > > >   LOG:    pid 19793:      watchdog: lifecheck started
> > > >
> > > > The tutorial would be helpful for you:
> > > >
> > http://www.pgpool.net/pgpool-web/contrib_docs/watchdog_master_slave/en.html
> > > >
> > > > >
> > > > > well I'm not asking for the watchdog timeout, thought it already
> > existed.
> > > > >
> > > > >
> > > > >
> > > > > On Sun, Apr 20, 2014 at 11:49 PM, Yugo Nagata <nagata at sraoss.co.jp>
> > > > wrote:
> > > > >
> > > > > > Hi,
> > > > > >
> > > > > > On Tue, 15 Apr 2014 16:00:50 -0300
> > > > > > Thomaz Luiz Santos <thomaz.santos at gmail.com> wrote:
> > > > > >
> > > > > > > I wonder if it is possible to configure the watchdog timeout,
> > I've
> > > > been
> > > > > > > doing some testing, and test the server timeout slave was waiting
> > > > for the
> > > > > > > master, the slave is to let take over as master if the master
> > does
> > > > not
> > > > > > > respond for a period of time . ?
> > > > > > >
> > > > > > > in the pgpool.conf
> > > > > > > health_check_timeout = 5
> > > > > >
> > > > > > health_check_timeout is for timeout of "health check", that is
> > about
> > > > > > backend PostgreSQL server.
> > > > > >
> > > > > >
> > > >
> > http://www.pgpool.net/docs/pgpool-II-3.2.8/doc/pgpool-en.html#HEALTH_CHECK_TIMEOUT
> > > > > >
> > > > > > In watchdog of 3.2, pgpool issues a query wd_lifecheck_query
> > ("SELECT
> > > > 1")
> > > > > > to
> > > > > > other pgpool every wd_interval seconds, and if this fails
> > wd_life_point
> > > > > > times,
> > > > > > pgpool regard the other as down.
> > > > > >
> > > >
> > http://www.pgpool.net/docs/pgpool-II-3.2.8/doc/pgpool-en.html#WD_INTERVAL
> > > > > >
> > > > > > But, parameter to configure timeout for watchdog doesn't exist.
> > > > > >
> > > > > > Are you requiring this?
> > > > > >
> > > > > > >
> > > > > > > but not work.
> > > > > > >
> > > > > > > pgpool 3.2.4
> > > > > > >
> > > > > > > --
> > > > > > > ------------------------------
> > > > > > > Thomaz Luiz Santos
> > > > > > > Linux User: #359356
> > > > > > > http://thomaz.santos.googlepages.com/
> > > > > >
> > > > > >
> > > > > > --
> > > > > > Yugo Nagata <nagata at sraoss.co.jp>
> > > > > >
> > > > >
> > > > >
> > > > >
> > > > > --
> > > > > ------------------------------
> > > > > Thomaz Luiz Santos
> > > > > Linux User: #359356
> > > > > http://thomaz.santos.googlepages.com/
> > > >
> > > >
> > > > --
> > > > Yugo Nagata <nagata at sraoss.co.jp>
> > > >
> > >
> > >
> > >
> > > --
> > > ------------------------------
> > > Thomaz Luiz Santos
> > > Linux User: #359356
> > > http://thomaz.santos.googlepages.com/
> >
> >
> > --
> > Yugo Nagata <nagata at sraoss.co.jp>
> >
> 
> 
> 
> -- 
> ------------------------------
> Thomaz Luiz Santos
> Linux User: #359356
> http://thomaz.santos.googlepages.com/


-- 
Yugo Nagata <nagata at sraoss.co.jp>


More information about the pgpool-general mailing list