SQL Cluster index problems

[JG5]

Hello,

I've been backing up a SQL Cluster now for a few months with no problems whatsoever until recently. On one of the nodes, when I try to restore, it shows only the dates that I did manual backups to choose from, while the other shows both. I've checked the index file location for the node in question and the dates contain therein span several months. The backup runs sucessfully every night, but I am unable to restore from dates other than manual backups.

Any ideas?

Thanks.

 

[wallace88]

You are using the NetWorker User GUI for SQL, right? At the bottom of the GUI, does it list the virtual node name as the client?

I presume when you say you have a SQL cluster, you mean that the SQL cluster is running off virtual node.

The virtual node should have it's own client index database. If it doesn't, it could be because NetWorker is updating the physical node's indexes.

If this is true, you need to create a file called pathownerignore, and put this file in the nsr\bin directory. This should be done on all the physical nodes.

In addition, you may have to edit the virtual node's client resource and change thebackup command so that it is:

nsrsqlsv -c <virtual_node_name>

Also make sure that the remote acces field contains the names of all the physical nodes in the cluster.

Once this is done, try backing up the virtual node again, then try restoring.

 

[JG5]

I am using the SQL module. Sure enough the node in question is listing the physical machine name rather than the virtual node name, however, both the virtual node and the physical node have their own indexes which are updated each night when the backup completes. I'm not familiar with the pathownerignore as you mentioned. Also, what does the -c switch do in the backup command field? I was under the impression that the -a switch should be used in older versions of the module to specify the virtual node. I haven't heard of the -c switch. My setup is:

2 Physical machines running 2 virtual nodes.

Thanks for the help.

 

[wallace88]

try listing the savesets for the virtual node with:

mminfo -av -ot -c <virtual node name>

Does it show the MSSQL: saveset? If it doesn't, then check the physical nodes.

pathownerignore is a zero length file that is used as a flag to NetWorker so that the files that is backed up during the virtual node backup does does not end up in the physical node's client file index (the physical node that was accessing the data at the time). This file should be in the nsr\bin subdirectory for all the physical nodes in the cluster.

What would you get if you started the NetWorker User GUI for SQL with a -s <server name> -a <virtual node name> option?