March 25, 2009 at 12:31 am
hi,
We have a/a/p clustered Sql server 2005.
Node 1 has one instance
Node 2 has 4 instances.
Node 3 is passive.
Now, our management decided to have DRT and for this we decided to go for Log shipping.I have couple of questions here.
1.We have total 32 databases in all 5 instances.Do we need to configure Logshipping for all 32 databases? and what about system databases in each instance?
2.In DRT location, can we install only one instance(Standalone) and logship these 32 databses to that instance?
3.If step 2 is possible, incase of disaster recovery how to point the applications? because we have 5 named instances and if we log ship all 32 databases to single instance in DRT location, how can I point the applications in case of DR
Please advice me and provide me the document(white paper) for configuring Log shipping
March 26, 2009 at 1:18 pm
Could you please give me some idea in configuring Logshiiping for this scenario?
is this possible? or Do I need to choose any otherway? please advice me..
thank you
March 27, 2009 at 8:10 am
We primarily use mirroring in our environments and we do not setup clusters in this way but I think this is certainly possible.
1. Yes log shipping is setup at the database level and will need to be setup for each database. I would not log ship system databases. MSDB actually houses tables that control log shipping so you couldn't log ship it, and master will of course contain things specific to each environment. If you have jobs, logins, linked servers, etc... that will need to be copied over then I would suggest building a custom process to keep environments in sync (This is out solution to it even with mirroring).
2. Yes you could log ship all the database from multiple instances to one stand alone instance. Log shipping even allows you to log ship to multiple instances.
3.If you are wanting to fail over Instances on a per node basis instead of on a cluster wide basis then i would suggest putting multiple network CNAMES in front of the cluster. All cnames would still point to the cluster, but your applications would use a specific cname depending on which node contained the databases they needed. Incase of a failover you would simply bring the databases on the DTR online and swing the CNAMES over to point to that box.
Makes sense?
MSDN has all the information you really need to setup log shipping.
March 27, 2009 at 8:58 am
I agree with Justin's advice. Just like mirroring, log shipping is by the database.
None of these technologies handles system databases. You have to handle those manually, probably periodic transfers of backups of these items.
Viewing 4 posts - 1 through 3 (of 3 total)
You must be logged in to reply to this topic. Login to reply