Master/Slave question

[Jason Dillon <Ja...@paybytouch.com>]

In a Master/Slave configuration, if my master goes down, I see that  
the Slave notices and starts accepting connections.

But if I start up the Master again, I notice that the slave does not  
transfer control back to the master.

If for some reason my Master node goes down, clients are configured  
to failover: to the client, and then operations restarts the  
master... do I now have to restart all clients to get them back on  
the master node, then restart the slave node to reestablish the  
master/slave configuration?

Ideally, I'd like if the master failed (say machine died due to power  
loss or something), that the slave would take over until it realized  
that the master was back on-line then resumed being a slave...  
redirecting clients back to the master after it has synchronized up  
the state between the two.

Is anything like this possible?

--jason
This email and any attachment(s) thereto, are intended for the use of
the addressee(s) named herein and may contain legally privileged and or
confidential information under applicable law. If you are not the
intended recipient of this e-mail, you are hereby notified any
dissemination, distribution or copying of this email, and any attachment(s)
thereto, is strictly prohibited. If you have received this communication
in error, please notify the sender at 415-281-2200 or via return e-mail at
postmaster@paybytouch.com and permanently delete the original copy and 
any copy of any e-mail, and any printout thereof. Thank you for your 
cooperation.

Re: Master/Slave question

[James Strachan <ja...@gmail.com>]

Basically we don't have an automatic synchronisation policy yet to
allow you to take an old master thats been down back up as a new
master; for now you have to manually stop the master & slave & move
the slaves data files over to the master then restart the master.

On 5/2/06, Javier Leyba <xl...@gmail.com> wrote:
> Jason
>
> I've asked the same some time ago and Rob told me the following:
>
> -----------
> master/slave doesn't handle this scenario - it is designed to provide
> a simple backup-support. When you wish to re-enable a master, you
> will need to ensure that the master is up to date with the slave -
> this is a manual process:
>
> 1. stop the slave
> 2. copy the data directories from the slave broker to the master.
> 3. re-start master then slave brokers.
> -------------------
>
> Hope this helps
>
> J
>
>
>
>
>
> On 5/2/06, Jason Dillon <Ja...@paybytouch.com> wrote:
> > In a Master/Slave configuration, if my master goes down, I see that
> > the Slave notices and starts accepting connections.
> >
> > But if I start up the Master again, I notice that the slave does not
> > transfer control back to the master.
> >
> > If for some reason my Master node goes down, clients are configured
> > to failover: to the client, and then operations restarts the
> > master... do I now have to restart all clients to get them back on
> > the master node, then restart the slave node to reestablish the
> > master/slave configuration?
> >
> > Ideally, I'd like if the master failed (say machine died due to power
> > loss or something), that the slave would take over until it realized
> > that the master was back on-line then resumed being a slave...
> > redirecting clients back to the master after it has synchronized up
> > the state between the two.
> >
> > Is anything like this possible?
>


--

James
-------
http://radio.weblogs.com/0112098/

Re: Master/Slave question

[Javier Leyba <xl...@gmail.com>]

Jason

I've asked the same some time ago and Rob told me the following:

-----------
master/slave doesn't handle this scenario - it is designed to provide
a simple backup-support. When you wish to re-enable a master, you
will need to ensure that the master is up to date with the slave -
this is a manual process:

1. stop the slave
2. copy the data directories from the slave broker to the master.
3. re-start master then slave brokers.
-------------------

Hope this helps

J





On 5/2/06, Jason Dillon <Ja...@paybytouch.com> wrote:
> In a Master/Slave configuration, if my master goes down, I see that
> the Slave notices and starts accepting connections.
>
> But if I start up the Master again, I notice that the slave does not
> transfer control back to the master.
>
> If for some reason my Master node goes down, clients are configured
> to failover: to the client, and then operations restarts the
> master... do I now have to restart all clients to get them back on
> the master node, then restart the slave node to reestablish the
> master/slave configuration?
>
> Ideally, I'd like if the master failed (say machine died due to power
> loss or something), that the slave would take over until it realized
> that the master was back on-line then resumed being a slave...
> redirecting clients back to the master after it has synchronized up
> the state between the two.
>
> Is anything like this possible?