[Linux-HA] heartbeat over ethernet with ipfail question

[Matthew Soffen]

That's why you have the serial port.  If all else fails, you'd have some
communication channel between the nodes.   You could have something else
( using Mon ? ) to monitor the network connection.

With the serial port, it would prevent both machines becoming master and
corrupting the disk ( since they would still be communicating to each
other ).  
 
What you do in an IP stack failure situation is up to you.  You could
monitor the ha logs or have something monitoring both nodes and killing
them if they are unable to communicate properly.  If the individual node
is not able to communicate to the other node via ether, it could reboot
the IP stack/etc.

Does this help ?

Matt




On Tue, 2005-03-15 at 10:41, R.Schramp at telecom.tno.nl wrote:
> Hello Matt,
> 
> this is exactly how I understood it from the manual,
> however I see not so much point in failing over to a system which can't
> use
> it's local IP stack (as in "can't ping the redundant default router").
> 
> Therefor in the case of not being able to ping, I would want the node to
> be/become slave.
> Is this possible?
> 
> I also have some doubts if drbd would successfully recover when both
> nodes loose each other for a while (ending up in 
> either Secondary/Unknown or Primary/Unknown)
> 
> Ruud
> 
> 
> > 
> > The problem is that with only an Ethernet communication line, 
> > you have a single point of failure ( the IP Stack ).  If 
> > something were to corrupt the IP stack on one machine (say, 
> > the standby), then it wouldn't be able to communicate to the 
> > master so it would take over.
> > 
> > Having 2 totally different lines of communication (i.e. a 
> > serial line ) you minimize that potential for problem because 
> > even if Ethernet croaks, the serial line would still be alive 
> > and communicating ( it would still see the master server ).
> > 
> > If that isn't possible, you might want to have a dedicated 
> > cross over Ethernet connection between the 2 servers.  That 
> > can mitigate some problems ( of the switches failing ).  You 
> > still need to deal with the single point of failure though ( 
> > the IP stack ).
> > 
> > Matt
> > 
> _______________________________________________
> Linux-HA mailing list
> Linux-HA at lists.linux-ha.org
> http://lists.linux-ha.org/mailman/listinfo/linux-ha