fuzzyocelot
Programmer
Hi everyone,
I'm confused and could really use some clarification on SQL Server 2005 failover clustering setup. The Windows Cluster has been installed, configured, and tested by one of our network admin guys. The O/S is Windows 2003 Server Enterprise SP2 and the servers are 64-bit. We currently have a SQL Server 2000 cluster running separately on 32-bit servers but that was setup long before I started working here. So setting up a cluster is new to me. BTW, we only have 2 nodes (active / passive). We need to install two SQL 2005 instances.
I’m currently reading “How to: Create a New SQL Server 2005 Failover Cluster (Setup)” in BOL and have some questions.
Under "Virtual SQL Server Name”, it says “the virtual SQL Sever name must be unique on your network, and must have a name that is different than the host cluster and cluster nodes”. The network guy gave us three node names. For example N1, N2, VN (not the real names). He said the cluster name is VN. So would that probably be the host cluster? Meaning, will we have to come up with a name that’s unique and different than those names for the virtual SQL server name?
Under “Virtual SQL Server Configuration”, it says “enter one IP address for each network configured for client access”. The network guy also gave us three IP addresses (two for the node names and one for the cluster name). Do I use these here or use the next available IP address or something completely different? I really don’t understand this part. Maybe it will make more sense once we start the installation.
Under “Service Accounts (Clusters)”, it says “the account for the SQL Server Agent service must be an administrator for the cluster node. Do not use the same account that was assigned to the Cluster service.” I don’t understand this. Can someone please explain why the SQL Agent service must be an admin for the cluster node? It doesn’t mention this under “Setting Up Windows Service Accounts” in BOL. I’m confused.
I also read that for failover clustering to work in SQL 2005, the service accounts for SQL server (including the SQL agent, etc.) MUST run as domain accounts that are members of a global or local domain group. We’re using Active Directory so we’ll have someone set up domain groups and accounts.
All the SQL services on the current SQL 2000 cluster use the same domain account. Is it really better to use separate accounts for each SQL service? We will be installing pretty much everything except for Full-Text Search. The article in BOL on “Setting Up Windows Service Accounts” lists the required Windows permissions for the database engine and the agent. However, it doesn’t list any Windows permissions for the other services. Does anyone know which Windows permissions these other services need?
I haven’t read the info on installing Analysis Services on a cluster yet. I’ll probably start a separate thread on that one, though, as I’m sure I’ll have a lot of questions.
I would really appreciate any help or advice!
Thank you!
I'm confused and could really use some clarification on SQL Server 2005 failover clustering setup. The Windows Cluster has been installed, configured, and tested by one of our network admin guys. The O/S is Windows 2003 Server Enterprise SP2 and the servers are 64-bit. We currently have a SQL Server 2000 cluster running separately on 32-bit servers but that was setup long before I started working here. So setting up a cluster is new to me. BTW, we only have 2 nodes (active / passive). We need to install two SQL 2005 instances.
I’m currently reading “How to: Create a New SQL Server 2005 Failover Cluster (Setup)” in BOL and have some questions.
Under "Virtual SQL Server Name”, it says “the virtual SQL Sever name must be unique on your network, and must have a name that is different than the host cluster and cluster nodes”. The network guy gave us three node names. For example N1, N2, VN (not the real names). He said the cluster name is VN. So would that probably be the host cluster? Meaning, will we have to come up with a name that’s unique and different than those names for the virtual SQL server name?
Under “Virtual SQL Server Configuration”, it says “enter one IP address for each network configured for client access”. The network guy also gave us three IP addresses (two for the node names and one for the cluster name). Do I use these here or use the next available IP address or something completely different? I really don’t understand this part. Maybe it will make more sense once we start the installation.
Under “Service Accounts (Clusters)”, it says “the account for the SQL Server Agent service must be an administrator for the cluster node. Do not use the same account that was assigned to the Cluster service.” I don’t understand this. Can someone please explain why the SQL Agent service must be an admin for the cluster node? It doesn’t mention this under “Setting Up Windows Service Accounts” in BOL. I’m confused.
I also read that for failover clustering to work in SQL 2005, the service accounts for SQL server (including the SQL agent, etc.) MUST run as domain accounts that are members of a global or local domain group. We’re using Active Directory so we’ll have someone set up domain groups and accounts.
All the SQL services on the current SQL 2000 cluster use the same domain account. Is it really better to use separate accounts for each SQL service? We will be installing pretty much everything except for Full-Text Search. The article in BOL on “Setting Up Windows Service Accounts” lists the required Windows permissions for the database engine and the agent. However, it doesn’t list any Windows permissions for the other services. Does anyone know which Windows permissions these other services need?
I haven’t read the info on installing Analysis Services on a cluster yet. I’ll probably start a separate thread on that one, though, as I’m sure I’ll have a lot of questions.
I would really appreciate any help or advice!
Thank you!