Disk exchange on two-node MS SQL 2000 active/active cluster

Hi!
I have two-node cluster on Windows 2000 Adv Server sp4. There are two
instanses of MS SQL 2000 Server sp3, one on each node. I have one
physical disk owened by the first node, that is not used. I need to
transfer it to the second node, name it with the letter this node
already has (removing the disk with this letter from the node by
renaming it or just by deleting) and place on it SQL Server logs that
were located on the disk with this letter.
This is my plan:
1. Transfer disk Z: from node SQL1 to node SQL2
2. Stop sql server and agent services on node SQL2 (or take them
offline with cluster administrator) WHich is better?
3. Copy log files from disk V: to disk Z: on node SQL2
4. Delete disk V: (or rename it as lets say P: ) - I plan to remove
the mirror with the former V: disk from storage
5. Rename disk Z: to disk V: on node SQL2
6. Start sql server and agent services on node SQL2 (or take them
online with cluster administrator)
Is the plan correct? Are there any dangers like (suppose) node signs
its drives and will not accept another disk because of it or because
the disks are a bit different in size? Anything else?
Thanks.
I will be doing something similar this Sunday morning, but with a 2003
cluster. Comments inline.
Geoff N. Hiten
Microsoft SQL Server MVP
Senior Database Administrator
Careerbuilder.com
I support the Professional Association for SQL Server
www.sqlpass.org
"Roust_m" <roustam@.hotbox.ru> wrote in message
news:a388fd78.0405210601.2d69da11@.posting.google.c om...
> Hi!
> I have two-node cluster on Windows 2000 Adv Server sp4. There are two
> instanses of MS SQL 2000 Server sp3, one on each node. I have one
> physical disk owened by the first node, that is not used. I need to
> transfer it to the second node, name it with the letter this node
> already has (removing the disk with this letter from the node by
> renaming it or just by deleting) and place on it SQL Server logs that
> were located on the disk with this letter.
> This is my plan:
> 1. Transfer disk Z: from node SQL1 to node SQL2
> 2. Stop sql server and agent services on node SQL2 (or take them
> offline with cluster administrator) WHich is better?
Use Cluster administrator to take the SQL service offline but leave the disk
resources online.
> 3. Copy log files from disk V: to disk Z: on node SQL2
>
Good.
> 4. Delete disk V: (or rename it as lets say P: ) - I plan to remove
> the mirror with the former V: disk from storage
>
Reletter it. Take it out of the SQL resource group.
> 5. Rename disk Z: to disk V: on node SQL2
>
Add this disk to the SQL resource group. Make SQL Server dependant on this
disk resource.
> 6. Start sql server and agent services on node SQL2 (or take them
> online with cluster administrator)
> Is the plan correct? Are there any dangers like (suppose) node signs
> its drives and will not accept another disk because of it or because
> the disks are a bit different in size? Anything else?
>
I have tested this once with a complete disk replacement using a test build.
Everything worked as expected. Make sure SQL stays offline while you are
moving things around. Make double sure you have fixed the dependancies.
Backup the universe before trying anything.
> Thanks.
|||Trying to transfer disk Z: from node SQL1 to node SQL2 (right mouse
button, change group, choose SQL2) get a window:
Are you sure the 'SQL1 Backup Z:' resource should be moved from the
''SQL1' group to the ''SQL2' group?
Choose "Yes"
Get another message:
Move Resources
The group will be changed for the following resources:
SQL1 Backup Z: (Only this resource, that means that I removed SQL
Server and dependancies from it)
Choose "Yes" again:
Get an error:
An error occured attempting to change the group of the 'Node2_BACKUP
Z:' resource to 'Node1'
The cluster node is not the owner of the group.
Error ID: 5016 (00001398).
Could not find how to change this.
"Geoff N. Hiten" <SRDBA@.Careerbuilder.com> wrote in message news:<urKIL3zPEHA.3012@.TK2MSFTNGP09.phx.gbl>...[vbcol=seagreen]
> I will be doing something similar this Sunday morning, but with a 2003
> cluster. Comments inline.
> --
> Geoff N. Hiten
> Microsoft SQL Server MVP
> Senior Database Administrator
> Careerbuilder.com
> I support the Professional Association for SQL Server
> www.sqlpass.org
> "Roust_m" <roustam@.hotbox.ru> wrote in message
> news:a388fd78.0405210601.2d69da11@.posting.google.c om...
> Use Cluster administrator to take the SQL service offline but leave the disk
> resources online.
> Good.
> Reletter it. Take it out of the SQL resource group.
> Add this disk to the SQL resource group. Make SQL Server dependant on this
> disk resource.
> I have tested this once with a complete disk replacement using a test build.
> Everything worked as expected. Make sure SQL stays offline while you are
> moving things around. Make double sure you have fixed the dependancies.
> Backup the universe before trying anything.
|||Bring both resource groups to the node you are logged into. Do the group
changes from that machine.
FYI, my changeover went smoothly. No issues with a replacement partition as
long as the group membership and dependancies are correct.
Geoff N. Hiten
Microsoft SQL Server MVP
Senior Database Administrator
Careerbuilder.com
I support the Professional Association for SQL Server
www.sqlpass.org
"Roust_m" <roustam@.hotbox.ru> wrote in message
news:a388fd78.0405220639.1bc712da@.posting.google.c om...
> Trying to transfer disk Z: from node SQL1 to node SQL2 (right mouse
> button, change group, choose SQL2) get a window:
> Are you sure the 'SQL1 Backup Z:' resource should be moved from the
> ''SQL1' group to the ''SQL2' group?
> Choose "Yes"
> Get another message:
> Move Resources
> The group will be changed for the following resources:
> SQL1 Backup Z: (Only this resource, that means that I removed SQL
> Server and dependancies from it)
> Choose "Yes" again:
> Get an error:
> An error occured attempting to change the group of the 'Node2_BACKUP
> Z:' resource to 'Node1'
> The cluster node is not the owner of the group.
> Error ID: 5016 (00001398).
> Could not find how to change this.
> "Geoff N. Hiten" <SRDBA@.Careerbuilder.com> wrote in message
news:<urKIL3zPEHA.3012@.TK2MSFTNGP09.phx.gbl>...[vbcol=seagreen]
disk[vbcol=seagreen]
this[vbcol=seagreen]
build.[vbcol=seagreen]
are[vbcol=seagreen]
|||Hi!
I tried to bring everything to the node I was loged on and the node I
was trying to give ownership of the disk (the same node) - the same
error. Just deleted this resource and recreated it on the correct node.
After this all went fine.
Thanks.
*** Sent via Developersdex http://www.codecomments.com ***
Don't just participate in USENET...get rewarded for it!