The node lost communication with cluster

Have you updated your NIC drivers?

Regards,
John

"Hasan" wrote in message

> I have a cluster with 3 nodes, under SYSTEM in the event
> viewer I receive those Warning message every 3 secondes.
>
>
> ID 1123:The node lost communication with cluster
> node 'NODE1' on network 'PublicA'.
> ID 1123:The node lost communication with cluster
> node 'NODE3' on network 'PublicA'.
> ID 1122:The node (re)established communication with
> cluster node 'Node1' on network 'PublicA'.
> ID 1122:The node (re)established communication with
> cluster node 'Node2' on network 'PublicA'.
>
>
> The cluster works well, the faileover works like it's
> suppose to do but we want to eliminate this tons of
> warning in the event viewer ...
>
>
> We have already change the adapter, the cable, the switch
> port, check the adapter config etc.... SO
>
>
> Thanks...
>
>
 >> Stay informed about: The node lost communication with cluster 

Clusters with more than two nodes use multi-cast messages to send
cluster-wide communications. Some switches have problems with multi-cast.
Then there is the issue of multi-cast addressing. You can tell there is an
issue when you get an error message mentioning the absence of a MADCAP
server. MADCAP is to multi-cast what DHCP is to single-cast IP. Read up
on multicast IP messaging and you will likely find how to tune your network
switches to avoid this problem.

--
Geoff N. Hiten
Senior Database Administrator
Microsoft SQL Server MVP




"Hasan" wrote in message

>I have a cluster with 3 nodes, under SYSTEM in the event
> viewer I receive those Warning message every 3 secondes.
>
>
> ID 1123:The node lost communication with cluster
> node 'NODE1' on network 'PublicA'.
> ID 1123:The node lost communication with cluster
> node 'NODE3' on network 'PublicA'.
> ID 1122:The node (re)established communication with
> cluster node 'Node1' on network 'PublicA'.
> ID 1122:The node (re)established communication with
> cluster node 'Node2' on network 'PublicA'.
>
>
> The cluster works well, the faileover works like it's
> suppose to do but we want to eliminate this tons of
> warning in the event viewer ...
>
>
> We have already change the adapter, the cable, the switch
> port, check the adapter config etc.... SO
>
>
> Thanks...
>
>
 >> Stay informed about: The node lost communication with cluster 

You may want to first verify that the Heartbeat Adapter is set at the top of
the intracluster communications list. It looks like the server is sending
heartbeats across the public network.

Shane

"Hasan" wrote:

> I have a cluster with 3 nodes, under SYSTEM in the event
> viewer I receive those Warning message every 3 secondes.
>
>
> ID 1123:The node lost communication with cluster
> node 'NODE1' on network 'PublicA'.
> ID 1123:The node lost communication with cluster
> node 'NODE3' on network 'PublicA'.
> ID 1122:The node (re)established communication with
> cluster node 'Node1' on network 'PublicA'.
> ID 1122:The node (re)established communication with
> cluster node 'Node2' on network 'PublicA'.
>
>
> The cluster works well, the faileover works like it's
> suppose to do but we want to eliminate this tons of
> warning in the event viewer ...
>
>
> We have already change the adapter, the cable, the switch
> port, check the adapter config etc.... SO
>
>
> Thanks...
>
>
 >> Stay informed about: The node lost communication with cluster 

I need some help on how to correct getting Event ID 1123 and Event ID 1122 on one of two cluster nodes every hour or some times frequently. Two times a not main node was removed from active cluster membership(Event ID 1135). This is a two node cluster running Windows Server 2003 EE SP1 (NovaScale 3045 Itanium2 )

In the event viewer every hour on one of two cluster nodes I receive such Warnings(one after another):
a)Event ID – 1123. The node lost communication with cluster node 'Cluster0' on network 'Heartbeat Cluster Connection'.
b) Event ID – 1123. The node lost communication with cluster node ' Cluster0' on network 'Public Network Connection'.

Or the other way

a) Event ID – 1123. The node lost communication with cluster node ' Cluster0' on network 'Public Network Connection'.
b)Event ID – 1123. The node lost communication with cluster node 'Cluster0' on network 'Heartbeat Cluster Connection'.


I have already read such articles:
• Recommended private "Heartbeat" configuration on a cluster server - http://support.microsoft.com/?id=258750
• Multicast support enabled for the cluster heartbeat http://support.microsoft.com/default.aspx?scid=kb;EN-US;307962

I have replaced Heartbeat connection on another pair of NIC‘s but still have same problem.

Settings of my Connections

The Heartbeat connection:
a) not using team adapter.
b) Internal cluster communication only (private only)
c) Cluster properties->Network priority: set on upper level
d) Network Connection->Advanced->Advanced Settings->Connections: not on upper level, after Private connection
The Public connection:
a) Using team adapter.
b) All comunications (mixed network)
c) Cluster properties->Network priority: after Heartbeat (not on upper level)
d) Network Connection->Advanced->Advanced Settings->Connections: set on upper level

May bee problem is in Heartbeat connection, or with node(which was removed from active cluster membership two times ) something wrong?

When I want to see Multicast configuration I type in CMD: cluster „ClusterName“ network „NetworkName“ /priv
And then I see empty columns - /T/Network/Name/Value/. Changing names of cluster and connection in command I have the same result. What does it means? The multicasting is enable or disable? And how to see corect Multicast settings?
I can disable multicasting (but there is important data and I afraid to lost it!) This is a correct way, or what shall I do?

May bee you can help me to eliminate this problem? Any suggestions?


From http://www.developmentnow.com/g/99_2006_2_0_0_698577/The-node-lost-com...ication

Posted via DevelopmentNow.com Groups
http://www.developmentnow.com
 >> Stay informed about: The node lost communication with cluster 

Three things you should look at:

1. You cannot use the public network for internal cluster communications
(heartbeat) because it is in a teaming configuration, even if you have a
dedicated heartbeat established. It is not supported. Switch the public
network for Client Only communication.
http://support.microsoft.com/?id=254101.

2. Make sure you have followed the heartbeat configuration best practices.
http://technet2.microsoft.com/windowsserver/en/library/8638a855-0ea3-4...-84d5-f
and
http://technet2.microsoft.com/windowsserver/en/library/68d8ec40-3cda-4...-bc13-4

3. Make sure the public network has been optimized for network optimization.
The corresponding registry entry
HKLM\SYSTEM\CurrentControlSet\Services\LanmanServer\Parameters Size value
should be set to 3. http://support.microsoft.com/kb/892422.

Sincerely,


Anthony Thomas, MCDBA, MCSA
Senior Consultant
Sogeti USA


"Denny" wrote in message

>I need some help on how to correct getting Event ID 1123 and Event ID 1122
>on one of two cluster nodes every hour or some times frequently. Two times
>a not main node was removed from active cluster membership(Event ID 1135).
>This is a two node cluster running Windows Server 2003 EE SP1 (NovaScale
>3045 Itanium2 )
>
> In the event viewer every hour on one of two cluster nodes I receive such
> Warnings(one after another):
> a)Event ID - 1123. The node lost communication with cluster node
> 'Cluster0' on network 'Heartbeat Cluster Connection'.
> b) Event ID - 1123. The node lost communication with cluster node '
> Cluster0' on network 'Public Network Connection'.
>
> Or the other way
>
> a) Event ID - 1123. The node lost communication with cluster node '
> Cluster0' on network 'Public Network Connection'.
> b)Event ID - 1123. The node lost communication with cluster node
> 'Cluster0' on network 'Heartbeat Cluster Connection'.
>
>
> I have already read such articles:
> . Recommended private "Heartbeat" configuration on a cluster server -
> http://support.microsoft.com/?id=258750
> . Multicast support enabled for the cluster heartbeat
> http://support.microsoft.com/default.aspx?scid=kb;EN-US;307962
>
> I have replaced Heartbeat connection on another pair of NIC's but still
> have same problem.
>
> Settings of my Connections
>
> The Heartbeat connection:
> a) not using team adapter.
> b) Internal cluster communication only (private only)
> c) Cluster properties->Network priority: set on upper level
> d) Network Connection->Advanced->Advanced Settings->Connections: not on
> upper level, after Private connection
> The Public connection:
> a) Using team adapter.
> b) All comunications (mixed network)
> c) Cluster properties->Network priority: after Heartbeat (not on upper
> level)
> d) Network Connection->Advanced->Advanced Settings->Connections: set on
> upper level
>
> May bee problem is in Heartbeat connection, or with node(which was removed
> from active cluster membership two times ) something wrong?
>
> When I want to see Multicast configuration I type in CMD: cluster
> "ClusterName" network "NetworkName" /priv
> And then I see empty columns - /T/Network/Name/Value/. Changing names of
> cluster and connection in command I have the same result. What does it
> means? The multicasting is enable or disable? And how to see corect
> Multicast settings?
> I can disable multicasting (but there is important data and I afraid to
> lost it!) This is a correct way, or what shall I do?
>
> May bee you can help me to eliminate this problem? Any suggestions?
>
>
> From
> http://www.developmentnow.com/g/99_2006_2_0_0_698577/The-node-lost-com...ication
>
> Posted via DevelopmentNow.com Groups
> http://www.developmentnow.com
 >> Stay informed about: The node lost communication with cluster