On Fri, Mar 30, 2007 at 11:13:48AM +0200, Max Hofer wrote:
> On Thursday 29 March 2007 22:32, Dejan Muhamedagic wrote:
> > On Thu, Mar 29, 2007 at 06:36:15PM +0200, Max Hofer wrote:
> > > I have a questiuon regarding the heartbeat messgage exchange.
> > > 
> > > Currently i have 2 cluster systems, each consisting of 2 node:
> > > - cluster A consists of nodes A1, A2
> > > - cluster B consists of nodes B1, B2
> > > 
> > > All 4 nodes are attached with bonded interface to a tow LAN
> > > switches SW1 and SW2 (lets call it normal LAN).
> > > 
> > > A1 and A2 (and B1 and B2) have a direct interconnection where
> > > the DRBD devices are syncronized plus a serial cable (lets call it
> > > DRBD LAN)
> > > 
> > > Thus currently cluster A (and B) use 3 different ways to exchange
> > > the heartbeat packages:
> > > - bcast ofer the DRDB LAN
> > > - ucast using normal lan
> > > - the serial cable
> > > I see 2 possible solutions:
> > > a) wrting a resource agent which polls the state from the other cluster
> > > and i use this state
> > 
> > Interesting idea. Not sure how tricky it would be to do right.
> > Depends also what for you would use that state. I guess to restart
> > some resources.
> My problem with this solution that i do not use the action-serialzing effect 
> of the transition engine.

I don't really get this one.

> > > b) i configure 1 single cib.xml with 2 "sub-clusters"
> > 
> > This is an obvious solution, but probably you'd have to do some
> > rewiring, i.e. all nodes should be equally well connected with
> > each other.
> This is out of question because i can  not interconnect all nodes to each
> other with a serial cable or with a directy LAN cable.

You could use a switch in case there are no security issues. I'm
afraid that they really have to be able to talk to each other.

> > > With sub-cluster i mean certain resource run only on cluster A and other
> > > resource run only on B.
> > 
> > I don't think that we have such a thing. The only option which
> > could be of help is to say that the cluster is asymmetrical, but
> > that won't buy you much either.
> This is already on place - resource location constraints based on node names.
> 
> > > My question now:
> > > * what will happen if one of the nodes is disconnected from the normal
> > > LAN - are the information tunnled over the redundant connections?
> > > 
> > > Scenario: A1 is disconnected from SW1. A2 still recieves HB packages 
> > > via the serial line and the DRBD LAN. Do B1 and B2 see A1 as dead or
> > > do the get the information about A1 via A2?
> > 
> > I don't think a node can get info about another node through a
> > third party. B1 and B2 will consider A1 gone. More, in this case,
> > they'd be a majority and the cluster would go on and reconsider
> > its options.
> My idea was A2 is attached to the LAN and the CIB change from B1/B2
> is propagated to A2. Then A2 propagates this change over the serical cable
> to A1.

CIB changes are always propagated. But all nodes have to be
connected.

> Maybe i really do not need this scenario and I probably just move all 
> mobile resources of cluster A from from A1 to A2.

Hmm, don't understand this one either. What's a mobile resource?

> I just wanted to know how this scenraio is handled by the heartbeat 
> framework.

Probably not very well.

> 
> kind regards
> Max

-- 
Dejan
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Reply via email to