February 1, 2006 at 8:53 am
I have a SQL job that runs for app. 46-72 hrs. It transfers a .dat file into a table. The server is rebooted once a week and can interuppt this job. Once the job begins again the program will error due to duplicate entries. How can I restart the program where it left off?
February 1, 2006 at 9:47 am
Hi - Just to give us a bit more to work on..
How are you running the transfer ?
Why is the server bounced once a week ?
If needed could you bounce it once the load is complete ?
February 1, 2006 at 12:13 pm
I'm doing the transfer through a DTS package in SQL 7. The job is set to run automatically on the 2nd of every month and the reboot is set automatically on Sunday at 12 pm every week. If the job collides with the reboot then the job will begin again after the server is back up. Since there are duplicates in the table then it causes an error and is a big problem due to how long the job takes to run.
February 2, 2006 at 3:28 am
If you are able to delete from the source you could try this: Batch up your records into bundles of say 10% of the total. Start a transaction insert a bundle delete the bundle from the source and then commit.
Another strategy might be to create a flag in the database that indicates if the load is runing or not. The re-boot process could check the flag and ony re-boot when the load is not running.
Viewing 4 posts - 1 through 4 (of 4 total)
You must be logged in to reply to this topic. Login to reply