![]() | |
![]() |
| | Thread Tools | Display Modes |
#1
| |||
| |||
|
|
Hello, We have a standalone Enterprise edition sql 2000 server that we want to migrate to a cluster and make it a virtual sql server, for high availability reasons. We do not want to change the name of the SQL Server, as hundreds of databases are on this server, and hundreds of applications refer to this SQL Server name. Because of these requirements, we cannot use the steps in BOL to upgrade our standalone to a cluster (that includes changing the sql server name) We are moving to two new servers (the current server will be retired) Does anyone know the following two questions.... Is it possible to take the system databases from a standalone version of SQL Server, and restore them to a clustered SQL Server? Are there differences in the system databases on a stand alone SQL Server vs. a clustered virtual SQL Server? Here is a high level idea of what we would like to do: - Stop SQL on the standalone server - Rename the standalone server - Install sql on the cluster, and give the virtual sql server the orignal name of my standalone server - shutdown sql on the cluster - copy the data and log files for all system and user databases to the cluster from the standalone server - start up sql on the cluster Thanks for your help and time! |
#2
| |||
| |||
|
|
Thank you for your comments and advice. I don't think the network name presented by the cluster should not be an issue, as long as the name of the virtual SQL Server instance is the same as before. The application users rely on the SQL Server name, not the cluster or server name. My main question is, are there any differences in the system databases on standalone servers vs. system databases of clustered virtual sql servers? I am having trouble finding the answer to that question. If someone can verify that for me I would very much appreciate it. ![]() Thank you! Kelly "Geoff N. Hiten" wrote: I wouldn't try and copy the system databases from one server to another, much less from a non-clustered instance to a clustered instance. You are likely much better off aliasing the old name to the new name using a DNS record or a client-side alias. Even if you do manage to copy the system databases, the network name presented to the outside world by the Cluster won't change so you will not solve your initial problem. Geoff N.Hiten Microsoft SQL Server MVP "KellyVV" <KellyVV (AT) discussions (DOT) microsoft.com> wrote in message news:6BC09EC5-ACC9-4493-8E89-43684AC9A82E (AT) microsoft (DOT) com... Hello, We have a standalone Enterprise edition sql 2000 server that we want to migrate to a cluster and make it a virtual sql server, for high availability reasons. We do not want to change the name of the SQL Server, as hundreds of databases are on this server, and hundreds of applications refer to this SQL Server name. Because of these requirements, we cannot use the steps in BOL to upgrade our standalone to a cluster (that includes changing the sql server name) We are moving to two new servers (the current server will be retired) Does anyone know the following two questions.... Is it possible to take the system databases from a standalone version of SQL Server, and restore them to a clustered SQL Server? Are there differences in the system databases on a stand alone SQL Server vs. a clustered virtual SQL Server? Here is a high level idea of what we would like to do: - Stop SQL on the standalone server - Rename the standalone server - Install sql on the cluster, and give the virtual sql server the orignal name of my standalone server - shutdown sql on the cluster - copy the data and log files for all system and user databases to the cluster from the standalone server - start up sql on the cluster Thanks for your help and time! |
#3
| |||
| |||
|
|
Hi Geoff, Yes, we definately plan on running some tests. I was hoping to get verification on whether or not the system databases are different before we tried it. If there are differences, that would end this option and I wouldn't need to run thru the test. One follow up question on your comment: "The network name presented to the outside world IS the Virtual SQL Instance name. Again, copying the system databases, even if successful, won't fix your problem." I plan on giving the Virtual SQL Instance name the same name as my current Stand alone SQL Server. That way the name of the SQL Server will not change. Let's say my standalone server is called "Server1". If I shut down sql on that server, and then rename that server to "Server2". Then I install SQL on my new cluster, and give the Virtual SQL Server the name "Server1". Won't that handle the name issues? Or am I missing something? Thanks. "Geoff N. Hiten" wrote: I am not 100% sure, but my gut tells me that the system database are different. The network name presented to the outside world IS the Virtual SQL Instance name. Again, copying the system databases, even if successful, won't fix your problem. You may want to build a test cluster using cheap recycled SCSI hardware or evena Virtual host (VMWare or Microsoft Virtual Server) as a training exercise. You can then break it as much as necessary without affecting your production systems. There is nothing like hands-on experience when you are planning a major server migration. Geoff N. Hiten Microsoft SQL Server MVP "KellyVV" <KellyVV (AT) discussions (DOT) microsoft.com> wrote in message news:E5B85548-C635-4161-8BE2-BADD5070E720 (AT) microsoft (DOT) com... Thank you for your comments and advice. I don't think the network name presented by the cluster should not be an issue, as long as the name of the virtual SQL Server instance is the same as before. The application users rely on the SQL Server name, not the cluster or server name. My main question is, are there any differences in the system databases on standalone servers vs. system databases of clustered virtual sql servers? I am having trouble finding the answer to that question. If someone can verify that for me I would very much appreciate it. ![]() Thank you! Kelly "Geoff N. Hiten" wrote: I wouldn't try and copy the system databases from one server to another, much less from a non-clustered instance to a clustered instance. You are likely much better off aliasing the old name to the new name using a DNS record or a client-side alias. Even if you do manage to copy the system databases, the network name presented to the outside world by the Cluster won't change so you will not solve your initial problem. Geoff N.Hiten Microsoft SQL Server MVP "KellyVV" <KellyVV (AT) discussions (DOT) microsoft.com> wrote in message news:6BC09EC5-ACC9-4493-8E89-43684AC9A82E (AT) microsoft (DOT) com... Hello, We have a standalone Enterprise edition sql 2000 server that we want to migrate to a cluster and make it a virtual sql server, for high availability reasons. We do not want to change the name of the SQL Server, as hundreds of databases are on this server, and hundreds of applications refer to this SQL Server name. Because of these requirements, we cannot use the steps in BOL to upgrade our standalone to a cluster (that includes changing the sql server name) We are moving to two new servers (the current server will be retired) Does anyone know the following two questions.... Is it possible to take the system databases from a standalone version of SQL Server, and restore them to a clustered SQL Server? Are there differences in the system databases on a stand alone SQL Server vs. a clustered virtual SQL Server? Here is a high level idea of what we would like to do: - Stop SQL on the standalone server - Rename the standalone server - Install sql on the cluster, and give the virtual sql server the orignal name of my standalone server - shutdown sql on the cluster - copy the data and log files for all system and user databases to the cluster from the standalone server - start up sql on the cluster Thanks for your help and time! |
#4
| |||
| |||
|
|
Hello, We have a standalone Enterprise edition sql 2000 server that we want to migrate to a cluster and make it a virtual sql server, for high availability reasons. We do not want to change the name of the SQL Server, as hundreds of databases are on this server, and hundreds of applications refer to this SQL Server name. Because of these requirements, we cannot use the steps in BOL to upgrade our standalone to a cluster (that includes changing the sql server name) We are moving to two new servers (the current server will be retired) Does anyone know the following two questions.... Is it possible to take the system databases from a standalone version of SQL Server, and restore them to a clustered SQL Server? Are there differences in the system databases on a stand alone SQL Server vs. a clustered virtual SQL Server? Here is a high level idea of what we would like to do: - Stop SQL on the standalone server - Rename the standalone server - Install sql on the cluster, and give the virtual sql server the orignal name of my standalone server - shutdown sql on the cluster - copy the data and log files for all system and user databases to the cluster from the standalone server - start up sql on the cluster Thanks for your help and time! |
#5
| |||
| |||
|
|
Clustered named instances show up as VirtualServerName\InstanceName. Since the Virtual Server and the Instance have the same name your system would start out as Server2\Server2. Renaming a clustered instance is possible but unsupported and highly unrecommended. How to change a clustered SQL Server Network Name http://support.microsoft.com/default...b;en-us;307336 Again, I strongly urge you to look at a DNS-based solution. http://www.faqs.org/rfcs/rfc2782.html GNH "KellyVV" <KellyVV (AT) discussions (DOT) microsoft.com> wrote in message news:61E53FEE-965B-4037-8C07-D60EC3DF5F54 (AT) microsoft (DOT) com... Hi Geoff, Yes, we definately plan on running some tests. I was hoping to get verification on whether or not the system databases are different before we tried it. If there are differences, that would end this option and I wouldn't need to run thru the test. One follow up question on your comment: "The network name presented to the outside world IS the Virtual SQL Instance name. Again, copying the system databases, even if successful, won't fix your problem." I plan on giving the Virtual SQL Instance name the same name as my current Stand alone SQL Server. That way the name of the SQL Server will not change. Let's say my standalone server is called "Server1". If I shut down sql on that server, and then rename that server to "Server2". Then I install SQL on my new cluster, and give the Virtual SQL Server the name "Server1". Won't that handle the name issues? Or am I missing something? Thanks. "Geoff N. Hiten" wrote: I am not 100% sure, but my gut tells me that the system database are different. The network name presented to the outside world IS the Virtual SQL Instance name. Again, copying the system databases, even if successful, won't fix your problem. You may want to build a test cluster using cheap recycled SCSI hardware or evena Virtual host (VMWare or Microsoft Virtual Server) as a training exercise. You can then break it as much as necessary without affecting your production systems. There is nothing like hands-on experience when you are planning a major server migration. Geoff N. Hiten Microsoft SQL Server MVP "KellyVV" <KellyVV (AT) discussions (DOT) microsoft.com> wrote in message news:E5B85548-C635-4161-8BE2-BADD5070E720 (AT) microsoft (DOT) com... Thank you for your comments and advice. I don't think the network name presented by the cluster should not be an issue, as long as the name of the virtual SQL Server instance is the same as before. The application users rely on the SQL Server name, not the cluster or server name. My main question is, are there any differences in the system databases on standalone servers vs. system databases of clustered virtual sql servers? I am having trouble finding the answer to that question. If someone can verify that for me I would very much appreciate it. ![]() Thank you! Kelly "Geoff N. Hiten" wrote: I wouldn't try and copy the system databases from one server to another, much less from a non-clustered instance to a clustered instance. You are likely much better off aliasing the old name to the new name using a DNS record or a client-side alias. Even if you do manage to copy the system databases, the network name presented to the outside world by the Cluster won't change so you will not solve your initial problem. Geoff N.Hiten Microsoft SQL Server MVP "KellyVV" <KellyVV (AT) discussions (DOT) microsoft.com> wrote in message news:6BC09EC5-ACC9-4493-8E89-43684AC9A82E (AT) microsoft (DOT) com... Hello, We have a standalone Enterprise edition sql 2000 server that we want to migrate to a cluster and make it a virtual sql server, for high availability reasons. We do not want to change the name of the SQL Server, as hundreds of databases are on this server, and hundreds of applications refer to this SQL Server name. Because of these requirements, we cannot use the steps in BOL to upgrade our standalone to a cluster (that includes changing the sql server name) We are moving to two new servers (the current server will be retired) Does anyone know the following two questions.... Is it possible to take the system databases from a standalone version of SQL Server, and restore them to a clustered SQL Server? Are there differences in the system databases on a stand alone SQL Server vs. a clustered virtual SQL Server? Here is a high level idea of what we would like to do: - Stop SQL on the standalone server - Rename the standalone server - Install sql on the cluster, and give the virtual sql server the orignal name of my standalone server - shutdown sql on the cluster - copy the data and log files for all system and user databases to the cluster from the standalone server - start up sql on the cluster Thanks for your help and time! |
#6
| |||
| |||
|
|
Thanks again Geoff. I wasn't clear on something - this is not a named instance of SQL. Both the current install and the Virutual SQL Server will be a default instance. I plan on naming the VIrtual SQL Server - 'Server1', which is the name of our current stand alone server. I will check out the links you sent. Thanks again. "Geoff N. Hiten" wrote: Clustered named instances show up as VirtualServerName\InstanceName. Since the Virtual Server and the Instance have the same name your system would start out as Server2\Server2. Renaming a clustered instance is possible but unsupported and highly unrecommended. How to change a clustered SQL Server Network Name http://support.microsoft.com/default...b;en-us;307336 Again, I strongly urge you to look at a DNS-based solution. http://www.faqs.org/rfcs/rfc2782.html GNH "KellyVV" <KellyVV (AT) discussions (DOT) microsoft.com> wrote in message news:61E53FEE-965B-4037-8C07-D60EC3DF5F54 (AT) microsoft (DOT) com... Hi Geoff, Yes, we definately plan on running some tests. I was hoping to get verification on whether or not the system databases are different before we tried it. If there are differences, that would end this option and I wouldn't need to run thru the test. One follow up question on your comment: "The network name presented to the outside world IS the Virtual SQL Instance name. Again, copying the system databases, even if successful, won't fix your problem." I plan on giving the Virtual SQL Instance name the same name as my current Stand alone SQL Server. That way the name of the SQL Server will not change. Let's say my standalone server is called "Server1". If I shut down sql on that server, and then rename that server to "Server2". Then I install SQL on my new cluster, and give the Virtual SQL Server the name "Server1". Won't that handle the name issues? Or am I missing something? Thanks. "Geoff N. Hiten" wrote: I am not 100% sure, but my gut tells me that the system database are different. The network name presented to the outside world IS the Virtual SQL Instance name. Again, copying the system databases, even if successful, won't fix your problem. You may want to build a test cluster using cheap recycled SCSI hardware or evena Virtual host (VMWare or Microsoft Virtual Server) as a training exercise. You can then break it as much as necessary without affecting your production systems. There is nothing like hands-on experience when you are planning a major server migration. Geoff N. Hiten Microsoft SQL Server MVP "KellyVV" <KellyVV (AT) discussions (DOT) microsoft.com> wrote in message news:E5B85548-C635-4161-8BE2-BADD5070E720 (AT) microsoft (DOT) com... Thank you for your comments and advice. I don't think the network name presented by the cluster should not be an issue, as long as the name of the virtual SQL Server instance is the same as before. The application users rely on the SQL Server name, not the cluster or server name. My main question is, are there any differences in the system databases on standalone servers vs. system databases of clustered virtual sql servers? I am having trouble finding the answer to that question. If someone can verify that for me I would very much appreciate it. ![]() Thank you! Kelly "Geoff N. Hiten" wrote: I wouldn't try and copy the system databases from one server to another, much less from a non-clustered instance to a clustered instance. You are likely much better off aliasing the old name to the new name using a DNS record or a client-side alias. Even if you do manage to copy the system databases, the network name presented to the outside world by the Cluster won't change so you will not solve your initial problem. Geoff N.Hiten Microsoft SQL Server MVP "KellyVV" <KellyVV (AT) discussions (DOT) microsoft.com> wrote in message news:6BC09EC5-ACC9-4493-8E89-43684AC9A82E (AT) microsoft (DOT) com... Hello, We have a standalone Enterprise edition sql 2000 server that we want to migrate to a cluster and make it a virtual sql server, for high availability reasons. We do not want to change the name of the SQL Server, as hundreds of databases are on this server, and hundreds of applications refer to this SQL Server name. Because of these requirements, we cannot use the steps in BOL to upgrade our standalone to a cluster (that includes changing the sql server name) We are moving to two new servers (the current server will be retired) Does anyone know the following two questions.... Is it possible to take the system databases from a standalone version of SQL Server, and restore them to a clustered SQL Server? Are there differences in the system databases on a stand alone SQL Server vs. a clustered virtual SQL Server? Here is a high level idea of what we would like to do: - Stop SQL on the standalone server - Rename the standalone server - Install sql on the cluster, and give the virtual sql server the orignal name of my standalone server - shutdown sql on the cluster - copy the data and log files for all system and user databases to the cluster from the standalone server - start up sql on the cluster Thanks for your help and time! |
#7
| |||
| |||
|
|
The default cluster SQL instance cannot be the same name as a host server. GNH "KellyVV" <KellyVV (AT) discussions (DOT) microsoft.com> wrote in message news:77BB2B41-EF4D-4CED-95C3-8DD1900574C2 (AT) microsoft (DOT) com... Thanks again Geoff. I wasn't clear on something - this is not a named instance of SQL. Both the current install and the Virutual SQL Server will be a default instance. I plan on naming the VIrtual SQL Server - 'Server1', which is the name of our current stand alone server. I will check out the links you sent. Thanks again. "Geoff N. Hiten" wrote: Clustered named instances show up as VirtualServerName\InstanceName. Since the Virtual Server and the Instance have the same name your system would start out as Server2\Server2. Renaming a clustered instance is possible but unsupported and highly unrecommended. How to change a clustered SQL Server Network Name http://support.microsoft.com/default...b;en-us;307336 Again, I strongly urge you to look at a DNS-based solution. http://www.faqs.org/rfcs/rfc2782.html GNH "KellyVV" <KellyVV (AT) discussions (DOT) microsoft.com> wrote in message news:61E53FEE-965B-4037-8C07-D60EC3DF5F54 (AT) microsoft (DOT) com... Hi Geoff, Yes, we definately plan on running some tests. I was hoping to get verification on whether or not the system databases are different before we tried it. If there are differences, that would end this option and I wouldn't need to run thru the test. One follow up question on your comment: "The network name presented to the outside world IS the Virtual SQL Instance name. Again, copying the system databases, even if successful, won't fix your problem." I plan on giving the Virtual SQL Instance name the same name as my current Stand alone SQL Server. That way the name of the SQL Server will not change. Let's say my standalone server is called "Server1". If I shut down sql on that server, and then rename that server to "Server2". Then I install SQL on my new cluster, and give the Virtual SQL Server the name "Server1". Won't that handle the name issues? Or am I missing something? Thanks. "Geoff N. Hiten" wrote: I am not 100% sure, but my gut tells me that the system database are different. The network name presented to the outside world IS the Virtual SQL Instance name. Again, copying the system databases, even if successful, won't fix your problem. You may want to build a test cluster using cheap recycled SCSI hardware or evena Virtual host (VMWare or Microsoft Virtual Server) as a training exercise. You can then break it as much as necessary without affecting your production systems. There is nothing like hands-on experience when you are planning a major server migration. Geoff N. Hiten Microsoft SQL Server MVP "KellyVV" <KellyVV (AT) discussions (DOT) microsoft.com> wrote in message news:E5B85548-C635-4161-8BE2-BADD5070E720 (AT) microsoft (DOT) com... Thank you for your comments and advice. I don't think the network name presented by the cluster should not be an issue, as long as the name of the virtual SQL Server instance is the same as before. The application users rely on the SQL Server name, not the cluster or server name. My main question is, are there any differences in the system databases on standalone servers vs. system databases of clustered virtual sql servers? I am having trouble finding the answer to that question. If someone can verify that for me I would very much appreciate it. ![]() Thank you! Kelly "Geoff N. Hiten" wrote: I wouldn't try and copy the system databases from one server to another, much less from a non-clustered instance to a clustered instance. You are likely much better off aliasing the old name to the new name using a DNS record or a client-side alias. Even if you do manage to copy the system databases, the network name presented to the outside world by the Cluster won't change so you will not solve your initial problem. Geoff N.Hiten Microsoft SQL Server MVP "KellyVV" <KellyVV (AT) discussions (DOT) microsoft.com> wrote in message news:6BC09EC5-ACC9-4493-8E89-43684AC9A82E (AT) microsoft (DOT) com... Hello, We have a standalone Enterprise edition sql 2000 server that we want to migrate to a cluster and make it a virtual sql server, for high availability reasons. We do not want to change the name of the SQL Server, as hundreds of databases are on this server, and hundreds of applications refer to this SQL Server name. Because of these requirements, we cannot use the steps in BOL to upgrade our standalone to a cluster (that includes changing the sql server name) We are moving to two new servers (the current server will be retired) Does anyone know the following two questions.... Is it possible to take the system databases from a standalone version of SQL Server, and restore them to a clustered SQL Server? Are there differences in the system databases on a stand alone SQL Server vs. a clustered virtual SQL Server? Here is a high level idea of what we would like to do: - Stop SQL on the standalone server - Rename the standalone server - Install sql on the cluster, and give the virtual sql server the orignal name of my standalone server - shutdown sql on the cluster - copy the data and log files for all system and user databases to the cluster from the standalone server - start up sql on the cluster Thanks for your help and time! |
#8
| |||
| |||
|
|
The ability to move the system and user databases between a standalone Enterprise instance and clustered Enterprise server instance as you are proposing is perfectly possible. I have not noticed any difference between clustered and non-clustered system databases, I however have not conducted any comprehensive comparison. The issues you will more likely encounter here are the same issues as migrating system databases between standalone instances of SQL. Have a look at 224071. http://support.microsoft.com/kb/224071. You may want to evaluate the possibility of migrating your user databases and specific system information (Logins etc) as an alternative to ensure your new setup clean. Your assumptions about the SQL instance names are correct, if I understand you correctly, are correct: You have a default standalone instance on ServerX (which will be removed from the network) which you wish to replace with a ServerX virtual server default instance. The only consideration here would be that there can not be another preexisting default instance of SQL installed on the cluster. Regards Gary Hope iSolve Business Solutions South Africa "KellyVV" <KellyVV (AT) discussions (DOT) microsoft.com> wrote in message news:6BC09EC5-ACC9-4493-8E89-43684AC9A82E (AT) microsoft (DOT) com... Hello, We have a standalone Enterprise edition sql 2000 server that we want to migrate to a cluster and make it a virtual sql server, for high availability reasons. We do not want to change the name of the SQL Server, as hundreds of databases are on this server, and hundreds of applications refer to this SQL Server name. Because of these requirements, we cannot use the steps in BOL to upgrade our standalone to a cluster (that includes changing the sql server name) We are moving to two new servers (the current server will be retired) Does anyone know the following two questions.... Is it possible to take the system databases from a standalone version of SQL Server, and restore them to a clustered SQL Server? Are there differences in the system databases on a stand alone SQL Server vs. a clustered virtual SQL Server? Here is a high level idea of what we would like to do: - Stop SQL on the standalone server - Rename the standalone server - Install sql on the cluster, and give the virtual sql server the orignal name of my standalone server - shutdown sql on the cluster - copy the data and log files for all system and user databases to the cluster from the standalone server - start up sql on the cluster Thanks for your help and time! |
![]() |
| Thread Tools | |
| Display Modes | |
| |