Hugo,

I think I saw a similar question recently in MSDN forum in T-SQL. I suggest you to ask there as well.

>Hi All,
>
>I have several scheduled tasks that run every 10-30 seconds. All this tasks run very well, they take in the worst cases 2 seconds to run, but usually is less than a second (History shows 00:00:00), when I run them in SQL they are almost instantaneous too, so everything is fine... Except there is one of this tasks, which connects to an Oracle Linked Server (there are two more that connect to the same oracle database) that once or twice a day decides to keep the query running ad-infinitum. Only way to stop it is to restart the SQL Server Agent, and then it runs smoothly for some time.
>
>My first thought to fix this problem is to limit the running time of the Jobs, lets say if they take more than 5 seconds then just cancel it and try again the next scheduled time, unfortunately I am not able to find information of how to do it (my google skills are not up to par this days it seems!), is it possible to do it? And how?
>
>I know it would be better to find the causes of this instead of treating the symptoms, but I thought it would be useful to have something like this anyways for all the scheduled jobs, so one can analyze the logs and when there are jobs that are "timing out" then one can try to find the problem without the need to stopping the job while investigating.
>
>
>Thanks
>
>Hugo