Cluster Service Hung on Starting

[Ayman Altounji]

I have recently installed Proliant G2 Cluster Package with MS 2000 Advanced Server SP2 with lastest compaq softpak 5.40. Whenever I start either node the server boots up normally all services start except for the cluster service. After about 4 mins from the last event in the system log I get an event 7022 - Cluster Service hung on starting and then about a minute later the Cluster Service starts with no errors and functions correctly.
If I stop and start the cluster service it takes about 1 min to start.
Why does the cluster service take so long to start on booting the servers.

Thanks for your help
Wayne

[Ayman Altounji]

This may be related to the network service or possibly a print spool. There are a couple of Microsoft Knowledge Base articles (Q258750 and Q319127) that you may want to check out (http://support.microsoft.com/). You should also search on the 7022 error message.

Network changes - if any network settings or changes to the network configuration are made after the cluster is setup then the service may be slow to start or not start at all. This can be caused by network binding order, name resolution, and RPC issues. MSCS uses remote procedure calls (RPC) and requires that the RPC service be running on both systems. Check to make sure that the RPC service is running on both systems and that the system event logs on each server do not have any RPC-related errors. Also, refer to Q-Article 258750 to ensure proper cluster network configuration settings.

There is a Q-article (Q319127) for NT 4.0 which states "Services Do Not Start and an Event ID 7022 Message Is Logged in the System Log of Event Viewer". This issue may occur if the Spooler service tries to start before the Server service. This issue may occur if you installed either the Lexmark International Markvision (LexBce) service or the Hewlett-Packard JetAdmin port monitoring software on your computer.

[Ayman Altounji]

One other thing that you might check is for virus scanning software. Most virus software wants to check the disks at boot time, and typically start before the Cluster Service. So when the Cluster Service tries to start, it can't access the disks because the virus software has them, so it waits. If you stop and start the Cluster Service, you won't have this probalem (and it will start faster) because the virus software has already checked the disks. So one possible solution is to try to delay the startup of your virus software until after the Cluster Service starts.

Thanks,
Doug

[Ayman Altounji]

Hi Doug
I have found exactly where the 3-4 min delay takes place by going through the cluster log file.
When the server boots and the cluster services begins to start there is a line in the cluster log:
Initializing Network Interfaces
Initialization Colmpete
The time between these two events is about 3-4 mins.
If however I start and stop the cluster service, the time between these two events is reduced to 1/100 of a second.
What is the cluster service doing when it tries to Initialize Network Interfaces.

Regards
Wayne

[Ayman Altounji]

I would double check your network settings - things like binding order, etc. Also, do you have NIC teaming enabled? I would also check the MS KB article (Q258750 ) to double check your network settings.

Thanks,
Doug

[Ayman Altounji]

Hi Doug
The Binding order for both interfaces on the server are:

File and Printer Sharing for Microsoft Networks
Internet Protocol (TCP/IP)
Client for Microsoft Networks
Internet Protocol (TCP/IP)

Is the order correct and can I do without either File and Printer Sharing for Microsoft Networks or Client for Microsoft Networks

Regards
Wayne

[Ayman Altounji]

Hi Doug
I finally solved the problem. I had Routing & Remote Access Service configured on both nodes in the cluster. I disabled this service and the cluster service started with no problems.

Thanks for all your help.
You pointed me in the right direction.

Wayne