SQL Scheduled backup jobs missing

[spiffy]

I have a SQL 2008 cluster on Windows Server 2008, running the 6.2.1.2 ba client and TSM for SQL 5.5.4.0.

I cannot get the TSM controlled schedule to run, it always shows as missing, yet i can launch the sqlfull.cmd and run the backup manually.

I have set up the Scheduler Service using the command line as recommended, the option files and sqlfull.log and sqlsched.log all reside on the same clustered drive (example E:\drive)

in the TSM Activity log, all i see is where the server is trying to prompt the scheduler to start a scheduled job, but it never starts.

I see nothing in any error logs, no errors about password prompts or the such, nothing in the activity log.. i am at a loss on what to do.

I have completely reconfigured the setup also by doing the following.

removed all cluster services, then removed the windows service for the scheduler, deleted all option files/log files/sqlfull.cmd, uninstalled teh client.

reinstalled the client on the host, recreated the option files and what not back into the clustered resource. Created the windows TSM scheduler service using the command line (set to manual startup), failed it over the other node, created teh service. launched the cluster administrator and created the cluster service resource.

And just to ensure i was not doing something wrong, i did the same steps on a SQL server cluster on Windows Server 2003, and it works, just not on Server 2008.

Any ideas?

 

[moon-buddy]

I don't want to ask the obvious but, do you have a separate TDP schedule that is Cluster defined for the TDP backup?

 

[spiffy]

yes, each clustered sql instance (there is 7 of them) has their own option file, log files, sqlfull.cmd and scheduler service, and it is set up in failover cluster administrator separatly too.

i have no problem setting this all up with Windows Server 2003 clusters, it just seems to be something with this 2008 cluster.

I have set the scheduler service to run using the account that has access to the sql instances, and this account is a full administrator (i can manually kick the sqlfull.cmd off when i log onto the server as this account, and it works)

i may need to open a PMR or something to do a trace and find out why the command is not launching..

this is the command i used to create the windows scheduler

dsmcutil install SCHEDuler /name:"TDPSQL Scheduler SQLCL02PRD01" /clientdir:"c:\Program Files\tivoli\tsm\baclient" /optfile:"E:\SQLCL02PRD01_tempDB_Logs_1\TSM\TDPSQL\dsm.opt" /node:SQLCL02PRD01 /passwordassword /validate:yes /autostart:no /startnow:no /clusternode:yes /clustername:NAMF219230

There are 7 more of these each pointing to a different optfile location on a different drive and each has a different nodename.

Oh my TSM server is 6.2.1.1

 

[moon-buddy]

What is /node:SQLCL02PRD01? Cluster node or control nodes?

 

[spiffy]

that is a cluster node. it is what i have set up on TSM as the nodename for this sql instance.

DSM.OPT File

NODename SQLCL02PRD01
DOMAIN E:
CLUSTERnode YES
PASSWORDAccess Generate

errorlogname "E:\SQLCL02PRD01_tempDB_Logs_1\TSM\TDPSQL\sqlerror.log"
schedlogname "E:\SQLCL02PRD01_tempDB_Logs_1\TSM\TDPSQL\sqlsched.log"


Here is the SQLFULL.CMD file that should launch the backup, and it does when manually executed

%sql_dir%\tdpsqlc backup * full /sqlserver=SQLCL02PRD01\SQLCL02PRD01 /tsmoptfile="E:\SQLCL02PRD01_tempdb_logs_1\tsm\tdpsql\dsm.opt" /logfile="E:\SQLCL02PRD01_tempdb_logs_1\tsm\tdpsql\sqlfull.log"

 

[hogmaster]

In similar solutions I have used polling mode.

 

[spiffy]

i will try the polling mode and see what happens from there.

 

[spiffy]

I figured this out. i needed to specify the TCPCLIENTADDRESS and TCPCLIENTPORT for each clusternode dsm.opt file

I had to use a unique tcpclientport for each one.

The prompted mode works now with these entries.