Hi Reader,
This is new to me: "blogging"; after some requests of a good friend and core-member of SQLUG.be, Bart ánd a lot of time of negotation with myself if I would find enough time to blog regularly, I decided to give it a try. In the about you can read about what my blog mainly will go about and why I choose the blogname "Por Dentro Sql".
Here we go with the first short posting:
Running on Sql Server 2005 Enterprise Edition, RTM x86 (but will be the same in SP1) I discovered that some jobs (all and only jobs created by maintenance plans) did run in the beginning and after some months they did not anylonger. However the SQLSERVERAGENT has run al the time. In the SQL Agent errorlog we saw "[000] Request to run job xxxx (from Schedule 5 (Daily)) refused because the job has been suspended".
I verified the job and the schedule; both were enabled.
Unless there have been already some restarts of the agent, the problem did continue. Since the reboot of the server that morning, the problem seems to disappear (the jobs where running fine again).
To have more insight into the problem, I tried to simulate the behaviour by trying to set a job suspended, without success.
I finally discovered: Sql Server Integration Services was not installed on the 2nd node. The reason is that SSIS (together with Notification Services and Reporting Services) are by default not cluster-aware and thus not installed (automatically) on all nodes (of course if you choose to install these services).
If you would, you can make SSIS cluster-aware as described in http://msdn2.microsoft.com/en-us/library/ms345193.aspx however, I decided to let them run locally.
Leason Learned: if you want to use Maintenance Plans (SSIS packages in general I think) always install (explicitly) SSIS on all nodes.
CU, PDS