Host unreachable in NLB and no heartbeats.
- From: Peter Werner <pwerner@xxxxxx>
- Date: Wed, 29 Jun 2005 23:59:25 +0200
Hello there,
I have configure a two node cluster at customer site like I had done before
in my test enviroment. We have two 2003 server running with two NICs each.
One NIC is in the subnet 192.168.60.0, the other NIC in 192.168.100.0. The
subnet 192.168.60.0 is the subnet from which the clients connect. Here I
have configured my Default Gateway and my DNS and also NLB. It looks like
that:
NODE 1
- NIC1: CIP 192.168.60.10
IP 192.168.60.4
DNS .........
GW .........
NLB
- NIC2: 192.168.100.4
NODE 2
- NIC1: CIP 192.168.60.10
IP 192.168.60.5
DNS .........
GW .........
NLB
- NIC2: 192.168.100.5
Default port rule with Multicast. Makes the second NIC unnecessary but this
still worked for me alright.
I can set up the NLB with the nlbmgr. I create the cluster, connect the
first host - no problem at all, and then the second, but when the cluster
tries to converge, an error is shown that NODE 2 is unreachable.
But I can ping that host, and I'm able to
select the network connection it the wizard to add the host!! So it can be
a problem with ICMP.
Interessting is that in the network analyzer not heartbeats are seen. I made
a comparision with a working NLB cluster in VMWare so I know what I should
see in the analyzer, but I don't see it.
Maybe some routs could be the problem? Because of other services, the rou-
ting table was adapted by my customer. But I really have no idea what to do.
We have changed the NICS - no success. If anybody has got an idea, I would
really appricate it.
:-)Peter
.
- Prev by Date: Clustered NFS shares
- Next by Date: Re: Cluster servers with different hardware configurations?
- Previous by thread: Clustered NFS shares
- Next by thread: How can I troubleshoot this ?
- Index(es):
Relevant Pages
|
