OS: Windows 2000 Server
SP level: 4 + all new hotfixes
HostMonitor: 4.10
Problem:
We have a cluster based on Windows 2000 Advanced Server with two nodes. We have some checks over the cluster name, mostly checking is a service is running. Quite sporadically we have errors checking those services.
Did anybody have this problem?
Any suggestion of what to do?
EDIT: just to be clear, the HostMonitor program is installed in another hardware (non-cluster) and the cluster is on the same physical LAN (100 Mbps with a CISCO ethernet switch).
Cluster checking (sometimes fail)
Yes,
I've similar problem on our Win2000 Cluster. (with exchange 2000)
Tests like UNC sometimes fails, without any significant reason. (also at night, when the load is very low) (via the node name, or via the cluster resource name)
When the test is bad, most of the time, it's for one cycle.
When I repeat the test immediatly after the failed test, the status returns to good.
I suppose the problem is a timing problem with MS clusters ?
I've changed the "consecutive bad results" parameter for those tests.
We have another two clusters, where I didn't see this problem.
I will check the difference with the other cluster. (Hardware, Servicepack, software, ...)
Regards,
Gert
I've similar problem on our Win2000 Cluster. (with exchange 2000)
Tests like UNC sometimes fails, without any significant reason. (also at night, when the load is very low) (via the node name, or via the cluster resource name)
When the test is bad, most of the time, it's for one cycle.
When I repeat the test immediatly after the failed test, the status returns to good.
I suppose the problem is a timing problem with MS clusters ?
I've changed the "consecutive bad results" parameter for those tests.
We have another two clusters, where I didn't see this problem.
I will check the difference with the other cluster. (Hardware, Servicepack, software, ...)
Regards,
Gert