Problem pinging other NLB servers...

Windows Server 2003 SP2 R2
Each server has two NICs: "Server NIC" and "NLB NIC"

I just added a new server to an existing NLB cluster. I've matched all the
settings on the new server to the existing NLB servers, however, since the
time I enabled and configured NLB on the new server it will no longer ping
the existing NLB servers in the cluster. The other NLB servers can ping the
new server.

When I viewed the arp table on the new server I could see that it was trying
to communicate over the "NLB NIC" interface to get to the other servers
because their was an arp entry on this interface. The original NLB servers
in the cluster do not have this problem and have arp entries under their
"Server NIC" interfaces and not their "NLB NIC" interfaces.

I then thought I would try to force the other NLB IPs to go through the
"Server NIC" by adding an entry to the routing table, and voila, it started
to respond to a ping. However, I don't understand why it is not working to
begin with. I shouldn't have to add the route.

Does anyone have any ideas why I'm having this problem and, is it OK to add
a persistant route to fix the problem?

Many Thanks!

Jaycee

Is there a chance that the labeling of the NIC's is reversed? where the NLB
NIC is really pluged into the server NIC ports/network and vise versa on the
other Server NIC. Usually it's not an issue to add an additional NLB server
to an existing group. And as to the routing addition, that is just a band-aid
covering the real issue.

Need more info to get further.
--
Chief Putz


"Jaycee" wrote:

> Windows Server 2003 SP2 R2
> Each server has two NICs: "Server NIC" and "NLB NIC"
>
> I just added a new server to an existing NLB cluster. I've matched all the
> settings on the new server to the existing NLB servers, however, since the
> time I enabled and configured NLB on the new server it will no longer ping
> the existing NLB servers in the cluster. The other NLB servers can ping the
> new server.
>
> When I viewed the arp table on the new server I could see that it was trying
> to communicate over the "NLB NIC" interface to get to the other servers
> because their was an arp entry on this interface. The original NLB servers
> in the cluster do not have this problem and have arp entries under their
> "Server NIC" interfaces and not their "NLB NIC" interfaces.
>
> I then thought I would try to force the other NLB IPs to go through the
> "Server NIC" by adding an entry to the routing table, and voila, it started
> to respond to a ping. However, I don't understand why it is not working to
> begin with. I shouldn't have to add the route.
>
> Does anyone have any ideas why I'm having this problem and, is it OK to add
> a persistant route to fix the problem?
>
> Many Thanks!
>
> Jaycee
>
>

I forgot to mention the hardware. The server is a Dell PE M600 (Blade
Server). The blade chassis houses 2 catalyst 3130 switches that the server
connects to. Both the NICs are on the same VLAN on the switches. What other
details would you need? You assistance is greatly appreciated.


"SharkMan" <> wrote in message
news:1E15BCFF-3872-42D8-8E03-...
> Is there a chance that the labeling of the NIC's is reversed? where the
> NLB
> NIC is really pluged into the server NIC ports/network and vise versa on
> the
> other Server NIC. Usually it's not an issue to add an additional NLB
> server
> to an existing group. And as to the routing addition, that is just a
> band-aid
> covering the real issue.
>
> Need more info to get further.
> --
> Chief Putz
>
>
> "Jaycee" wrote:
>
>> Windows Server 2003 SP2 R2
>> Each server has two NICs: "Server NIC" and "NLB NIC"
>>
>> I just added a new server to an existing NLB cluster. I've matched all
>> the
>> settings on the new server to the existing NLB servers, however, since
>> the
>> time I enabled and configured NLB on the new server it will no longer
>> ping
>> the existing NLB servers in the cluster. The other NLB servers can ping
>> the
>> new server.
>>
>> When I viewed the arp table on the new server I could see that it was
>> trying
>> to communicate over the "NLB NIC" interface to get to the other servers
>> because their was an arp entry on this interface. The original NLB
>> servers
>> in the cluster do not have this problem and have arp entries under their
>> "Server NIC" interfaces and not their "NLB NIC" interfaces.
>>
>> I then thought I would try to force the other NLB IPs to go through the
>> "Server NIC" by adding an entry to the routing table, and voila, it
>> started
>> to respond to a ping. However, I don't understand why it is not working
>> to
>> begin with. I shouldn't have to add the route.
>>
>> Does anyone have any ideas why I'm having this problem and, is it OK to
>> add
>> a persistant route to fix the problem?
>>
>> Many Thanks!
>>
>> Jaycee
>>
>>

What does "route" at a cmd prompt show for your default/main gateway? I'm
guessing since your traffic is running out your nlb nic then it see's that as
your default, I would think you need to change the default to your corp nic
as one option. next you mentioned that you used the new hosts config util to
join the existing nlb group, I haev had some issues that way, I now setup the
basic nic stuctture on the new host but then use one of the existing nlb
hosts to join it to the group (they won't be able to do it if any thing is
wrong on the new host).

--
Chief Putz


"Jaycee" wrote:

> I forgot to mention the hardware. The server is a Dell PE M600 (Blade
> Server). The blade chassis houses 2 catalyst 3130 switches that the server
> connects to. Both the NICs are on the same VLAN on the switches. What other
> details would you need? You assistance is greatly appreciated.
>
>
> "SharkMan" <> wrote in message
> news:1E15BCFF-3872-42D8-8E03-...
> > Is there a chance that the labeling of the NIC's is reversed? where the
> > NLB
> > NIC is really pluged into the server NIC ports/network and vise versa on
> > the
> > other Server NIC. Usually it's not an issue to add an additional NLB
> > server
> > to an existing group. And as to the routing addition, that is just a
> > band-aid
> > covering the real issue.
> >
> > Need more info to get further.
> > --
> > Chief Putz
> >
> >
> > "Jaycee" wrote:
> >
> >> Windows Server 2003 SP2 R2
> >> Each server has two NICs: "Server NIC" and "NLB NIC"
> >>
> >> I just added a new server to an existing NLB cluster. I've matched all
> >> the
> >> settings on the new server to the existing NLB servers, however, since
> >> the
> >> time I enabled and configured NLB on the new server it will no longer
> >> ping
> >> the existing NLB servers in the cluster. The other NLB servers can ping
> >> the
> >> new server.
> >>
> >> When I viewed the arp table on the new server I could see that it was
> >> trying
> >> to communicate over the "NLB NIC" interface to get to the other servers
> >> because their was an arp entry on this interface. The original NLB
> >> servers
> >> in the cluster do not have this problem and have arp entries under their
> >> "Server NIC" interfaces and not their "NLB NIC" interfaces.
> >>
> >> I then thought I would try to force the other NLB IPs to go through the
> >> "Server NIC" by adding an entry to the routing table, and voila, it
> >> started
> >> to respond to a ping. However, I don't understand why it is not working
> >> to
> >> begin with. I shouldn't have to add the route.
> >>
> >> Does anyone have any ideas why I'm having this problem and, is it OK to
> >> add
> >> a persistant route to fix the problem?
> >>
> >> Many Thanks!
> >>
> >> Jaycee
> >>
> >>
>

I finally got it working correctly. The Dell tech that I spoke with said
that the hardware and OS may detect differently which card is the "primary"
NIC. The "Server NIC" was configured on the second NIC in Windows and the
"NLB NIC" was on the first NIC in Windows. I swapped the configs on both of
these and ensured that the network bindings under the advanced network
settings were in the correct order. After a reboot everything then worked
when I configured NLB.

Thanks!

"SharkMan" <> wrote in message
news:4A955397-F9A7-41F4-B9A5-...
> What does "route" at a cmd prompt show for your default/main gateway? I'm
> guessing since your traffic is running out your nlb nic then it see's that
> as
> your default, I would think you need to change the default to your corp
> nic
> as one option. next you mentioned that you used the new hosts config util
> to
> join the existing nlb group, I haev had some issues that way, I now setup
> the
> basic nic stuctture on the new host but then use one of the existing nlb
> hosts to join it to the group (they won't be able to do it if any thing is
> wrong on the new host).
>
> --
> Chief Putz
>
>
> "Jaycee" wrote:
>
>> I forgot to mention the hardware. The server is a Dell PE M600 (Blade
>> Server). The blade chassis houses 2 catalyst 3130 switches that the
>> server
>> connects to. Both the NICs are on the same VLAN on the switches. What
>> other
>> details would you need? You assistance is greatly appreciated.
>>
>>
>> "SharkMan" <> wrote in message
>> news:1E15BCFF-3872-42D8-8E03-...
>> > Is there a chance that the labeling of the NIC's is reversed? where the
>> > NLB
>> > NIC is really pluged into the server NIC ports/network and vise versa
>> > on
>> > the
>> > other Server NIC. Usually it's not an issue to add an additional NLB
>> > server
>> > to an existing group. And as to the routing addition, that is just a
>> > band-aid
>> > covering the real issue.
>> >
>> > Need more info to get further.
>> > --
>> > Chief Putz
>> >
>> >
>> > "Jaycee" wrote:
>> >
>> >> Windows Server 2003 SP2 R2
>> >> Each server has two NICs: "Server NIC" and "NLB NIC"
>> >>
>> >> I just added a new server to an existing NLB cluster. I've matched
>> >> all
>> >> the
>> >> settings on the new server to the existing NLB servers, however, since
>> >> the
>> >> time I enabled and configured NLB on the new server it will no longer
>> >> ping
>> >> the existing NLB servers in the cluster. The other NLB servers can
>> >> ping
>> >> the
>> >> new server.
>> >>
>> >> When I viewed the arp table on the new server I could see that it was
>> >> trying
>> >> to communicate over the "NLB NIC" interface to get to the other
>> >> servers
>> >> because their was an arp entry on this interface. The original NLB
>> >> servers
>> >> in the cluster do not have this problem and have arp entries under
>> >> their
>> >> "Server NIC" interfaces and not their "NLB NIC" interfaces.
>> >>
>> >> I then thought I would try to force the other NLB IPs to go through
>> >> the
>> >> "Server NIC" by adding an entry to the routing table, and voila, it
>> >> started
>> >> to respond to a ping. However, I don't understand why it is not
>> >> working
>> >> to
>> >> begin with. I shouldn't have to add the route.
>> >>
>> >> Does anyone have any ideas why I'm having this problem and, is it OK
>> >> to
>> >> add
>> >> a persistant route to fix the problem?
>> >>
>> >> Many Thanks!
>> >>
>> >> Jaycee
>> >>
>> >>
>>