All,
We have a number of Windows 2000 Advanced Servers that are set-up to use
network load balancing for resilience. However, some of the servers upon
reboot will sometimes not re-join the cluster - instead they just say they
are converging. I have no idea why. Most of the time a reboot will fix this,
other times I have to remove and reinstall the Network Load Balancing
component.
Any ideas why this may be happening?
More details:
E.g. There are 8 hosts and host number 7 reboots. After reboot I do a 'wlbs
query' and get the following:
Host 7 converging with the following host(s) as part of the cluster:
1, 2, 3, 4, 5, 6, 8
Notice that host 7 isn't actually listed along the second line with the rest
of the servers. If I do the same command on the rest of the servers they
also show all hosts except for host 7, and show that they are converged. The
cluster also functions perfectly fine without host 7.
Yet if I restart server 7 again, it will successfully re-join the cluster
around 9 out of 10 times. The times it doesn't re-join after a second
reboot, I have to remove the Network Load Balancing component, re-boot,
re-add the Network Load balancing component, configure it, re-boot again and
all is usually fine.
The only entry in event viewer is the usual information entry WLBS : cluster
mode started with host ID 7.
One interesting point is that usually only hosts 6, 7 and 8 have this
problem. Hosts 1 to 5 are perfectly fine.
Hosts 1 to 5 are new builds and were the first servers in the cluster. Hosts
6, 7 and 8 were used in a different cluster on a completely different subnet
around a year ago. While in the previous cluster and subnet, they worked
perfectly fine and we did not have this problem. The problem only started
since installing them in the new cluster/subnet. All NIC's have had their IP
addresses/masks changed and all static routes have been deleted/corrected
for the new subnet before introducing them into the new cluster.
What am I missing?
Cheers in advance,
Paul.
|
Similar Threads
Linear Mode
