dejanmm at fastmail
Nov 2, 2009, 2:53 AM
Post #5 of 5
(201 views)
Permalink
|
Hi,
On Mon, Oct 26, 2009 at 12:53:32PM +0800, Joe P.H. Chiang wrote:
> Hi Thank you for your reply
> this video may show the better idea with configuration and video of the 3
> nodes and the problem i have..
> http://www.youtube.com/watch?v=jCCqRGF90RE
Watching videos is great, but there's not much substance in it :)
> When lb1 was suspended via vmware
> immediately the lb2 and lb3 start to listen to
> lb1's virtual ip address
That's confusing. A single (not cloned) resource instance should
run only on one node.
> what i want is that only lb3 listens to lb1's virtual ip
>
> and then when lb3 is goes down(suspended via vmware) the lb2 to listen lb1's
> virtual ip which is 192.168.19.101
>
> and how do i do that?
By using location constraints (node preferences).
Can you please post your configuration and the logs. You can use
hb_report.
Thanks,
Dejan
>
> Problem:
> When load balancer 1 is down then the load balancer 2 and load balancer 3
> automatically listen to virtual IP of load balancer 1
>
> Setup: 3 Nodes, 2 Active 1 Backup
> Load Balancer 1= Front end
> Load Balancer 2= Front end
> Load Balanccr 3= Backup
>
> lb1 Virtual IP = 192.168.19.101
> lb2 Virtual IP = 192.168.19.102
>
> lb1 Physical IP = 192.168.19.128
> lb2 Physical IP = 192.168.19.135
> lb3 Physical IP = 192.168.19.132
>
> I hope you guys can understand my poor English this time
> Thanks
>
> Joe
>
>
>
>
> On Fri, Oct 23, 2009 at 10:42 PM, Dejan Muhamedagic <dejanmm[at]fastmail.fm>wrote:
>
> > Hi,
> >
> > On Fri, Oct 23, 2009 at 07:42:05PM +0800, Joe P.H. Chiang wrote:
> > > I have some trouble setting this up.
> > >
> > > I have 3 Nodes
> > > LB1 : listen to 10.10.10.1
> > > LB2 : listen to 10.10.10.2
> > > LB3 : backup node for lb1 & lb2
> > >
> > >
> > > Setup1
> > > if LB1 goes DOWN.. LB3(backup node) goes UP then, if LB3(backup node)
> > DOWN
> > > then LB2 will host (LB1+LB2) eth0:0 = 10.10.10.1 + eth0:1= 10.10.10.2
> >
> > Not sure what do you mean with goes up/down.
> >
> > > same goes for LB2
> > >
> > > Setup2
> > > if LB2 DOWN.. LB3(backup node) goes UP then, if LB3(backup node) DOWN
> > then
> > > LB1 will host (LB1+LB2) eth0:0 = 10.10.10.2 + eth0:1= 10.10.10.1
> > >
> > > Problem: Setup1
> > > when LB1 goes DOWN then LB3 and LB2 will BOTH host the IP of LB1.
> > > or
> > > when LB2 goes DOWN then LB1 and LB3 will BOTH host the IP of LB2
> >
> > You mean one non-cloned resource runs on two nodes? That
> > shouldn't happen. It can happen on split-brain and quorum
> > ignored.
> >
> > > hmm is there any reason why?
> > >
> > > kinda hard to explain this..
> >
> > Kinda hard to help without logs and configuration.
> >
> > Thanks,
> >
> > Dejan
> >
> > > _______________________________________________
> > > Linux-HA mailing list
> > > Linux-HA[at]lists.linux-ha.org
> > > http://lists.linux-ha.org/mailman/listinfo/linux-ha
> > > See also: http://linux-ha.org/ReportingProblems
> > _______________________________________________
> > Linux-HA mailing list
> > Linux-HA[at]lists.linux-ha.org
> > http://lists.linux-ha.org/mailman/listinfo/linux-ha
> > See also: http://linux-ha.org/ReportingProblems
> >
>
>
>
> --
> Thanks,
> Joe
> _______________________________________________
> Linux-HA mailing list
> Linux-HA[at]lists.linux-ha.org
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems
_______________________________________________
Linux-HA mailing list
Linux-HA[at]lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
|
