![]() | |
![]() |
| | Thread Tools | Display Modes |
#1
| |||
| |||
|
#2
| |||
| |||
|
|
I'm moving our production DB from a stand alone install to an active/passive cluster. Current system is running Windows 2008 and SQL 2008 SP1, the cluster will run Windows 2008 R2. The machine that is actively running the database now will become the second node in the cluster when it's all said and done. I think have two options and am looking for any advice or other ideas. Option 1: Setup a single node cluster and restore the system databases. When ready to migrate, stop SQL on the existing server and change the LUNs over to the new server. Attach the databases and things should be ready to go. Then ugrade the old server and bring it in as the second node. Option 2: Setup a single node cluster using newly provisioned LUNs that match the current production setup. Backup system and user databases from the current server and restore them onto the cluster. Then ugrade the old server and bring it in as the second node. Option 1 seems much easier, Option 2 seems safer but more work. I have a lab where I've been testing this stuff, but curious what others think. |
#3
| |||
| |||
|
|
Hi Phavel, According to me. I feel there may be risk in option 1 & option 2 because the server name of the old system and new system will be different. You may have to find out the additional entries required to be configured or migrated for cluster and non-cluster system. The best option may be 1. Install single node cluster 2. Migrate Logins & SQL Server jobs 3. Restore user databases. 4. Rebuild the old server and attach it to the cluster. Regards, Balaji "phavel" <phavel (AT) discussions (DOT) microsoft.com> wrote in message news:75F8ABD3-EF0A-4ABF-83BA-9762B1B91852 (AT) microsoft (DOT) com... I'm moving our production DB from a stand alone install to an active/passive cluster. Current system is running Windows 2008 and SQL 2008 SP1, the cluster will run Windows 2008 R2. The machine that is actively running the database now will become the second node in the cluster when it's all said and done. I think have two options and am looking for any advice or other ideas. Option 1: Setup a single node cluster and restore the system databases. When ready to migrate, stop SQL on the existing server and change the LUNs over to the new server. Attach the databases and things should be ready to go. Then ugrade the old server and bring it in as the second node. Option 2: Setup a single node cluster using newly provisioned LUNs that match the current production setup. Backup system and user databases from the current server and restore them onto the cluster. Then ugrade the old server and bring it in as the second node. Option 1 seems much easier, Option 2 seems safer but more work. I have a lab where I've been testing this stuff, but curious what others think. . |
#4
| |||
| |||
|
|
Thanks, I've not had trouble in the lab with option 2, but you make a good point about possible inconsistencies. Do any of the Microsoft folks know if there is a technet or msdn article on the proper procedure for doing this type of migration? Thanks. "Balaji" wrote: Hi Phavel, According to me. I feel there may be risk in option 1 & option 2 because the server name of the old system and new system will be different. You may have to find out the additional entries required to be configured or migrated for cluster and non-cluster system. The best option may be 1. Install single node cluster 2. Migrate Logins & SQL Server jobs 3. Restore user databases. 4. Rebuild the old server and attach it to the cluster. Regards, Balaji "phavel" <phavel (AT) discussions (DOT) microsoft.com> wrote in message news:75F8ABD3-EF0A-4ABF-83BA-9762B1B91852 (AT) microsoft (DOT) com... I'm moving our production DB from a stand alone install to an active/passive cluster. Current system is running Windows 2008 and SQL 2008 SP1, the cluster will run Windows 2008 R2. The machine that is actively running the database now will become the second node in the cluster when it's all said and done. I think have two options and am looking for any advice or other ideas. Option 1: Setup a single node cluster and restore the system databases. When ready to migrate, stop SQL on the existing server and change the LUNs over to the new server. Attach the databases and things should be ready to go. Then ugrade the old server and bring it in as the second node. Option 2: Setup a single node cluster using newly provisioned LUNs that match the current production setup. Backup system and user databases from the current server and restore them onto the cluster. Then ugrade the old server and bring it in as the second node. Option 1 seems much easier, Option 2 seems safer but more work. I have a lab where I've been testing this stuff, but curious what others think. . |
![]() |
| Thread Tools | |
| Display Modes | |
| |