Configuring MS-SQL backup

[bst740]

Hello everyone!

I have 2 question on configuring MS-SQL Backup

1) I created 2 scripts in client: first for full backup of my databases, second for transaction logs backup for the sames databases. Can I Configure it in one policy? I need to execute full backup once at week, and transaction log once at hour. How do I need to configure policy and schedules? Do I need to create a new policy for transaction logs?

2) My backup executes correctly, when it's initiated by master server during the backup window. But I can't execute manual backup, it fails. How to configure it correctly?

Thanks!

P.S. Master, media, client are 7.5.0.6 running on Win 2008 R2, MS SQL 2008

[Mark_Solutions]

The log doesn't seem to go as far as the failure

Have a look for a log under \netbackup\logs\mssql_backup_failures\ to see if we can see more information

Also, if that is the logs backup please show us the bch file you used for that one

As this is a cluster make sure you have configured the sql gui on both nodes and have the bch files created on both nodes to allow for failovers.

Thanks

[bst740]

Log backup and msssql_backup failures are in attach

SQL GUI is configured on both nodes, all scripts are on shared disk, this should not be a problem

Thanks!

[Mark_Solutions]

In that log it says "CLIENT VDC01-DISQL--N1"

Is the policy set to use the cluster name?

What are the node names (I am guessing CLIENT VDC01-DISQL--N1 and CLIENT VDC01-DISQL--N2 ?)

The script looks OK so I am wondering what it cannot communicate with - maybe you also need bpcd logging setup on the client too to see what it is doing wrong.

[bst740]

Mark, I don't know why, but now backup with test script goes normally without errors. I didn't chang anything. I only turned on bpcd log (in attach). Maybe I need to wait a while to make sure that everything is normal

VDC01... - are a names of nodes, CDC01 is a name of cluster

In policy, on BAR and SQL client I configure only cluster name

[Mark_Solutions]

It will take a little while after the full backup completes for SQL to do its post backup processing so that could have been the issue

You should now be able to edit the bch file to use the $ALL or the specific database names that you want

Always do the full backup first

When you do the log backups check the logs if you get a status 1 as some may need their mode changing in order to do log backups

Glad things are looking sorted for you

[bst740]

Unfortunately, the error 48 appears regularly again

client hostname could not be found(48)

04.09.2013 12:59:48 - Error bprd(pid=4896) CLIENT CDC01-DISQL--01.DC-DEVELOP.TN.CORP  POLICY DI_SQL_logs  SCHED NONE  EXIT STATUS 48 (client hostname could not be found)

Errors for full backupa are the same

I don't now what to check

:(

[Mark_Solutions]

In the bpcd log it seems to be resolving the Master and Media Servers as IP addresses only

On the nodes of the SQL Cluster add Hosts file entries for the Master and Media Server using their short and FQDN

Being a cluster I guess it also has multiple networks and so using hosts files all round (ones on the Master and Media Servers with entries for the nodes and virtual name of the SQL Server) may also help

Try adding these and try again - the fact that it sometimes works and sometimes doesn't may indicate that it sometimes uses the wromng network and seeing the connection come in as an IP address only also needs addressing