March 21, 2006 at 1:33 pm
We have several Oracle info servers set up as Linked servers to our SQL Server. We are experiencing intermittent problems when a query is executed against a linked server The server reports a 7399 error. Once the error is received any query against any oracle linked server we have will return that same error on that SPID . I can close my instance of QA wait several minutes and try again and the result set returns as expected most of the time.
We have a job set to loop through the Linked servers overnight to collect switch data for our reports. Once the error is received all the remaining steps will return the same error. The error is causing our automated processing to fail. These are jobs and stored procedures that have been in place for quite some time and nothing has changed as far as our dB is concerned.
I have tried drop and readd with no success. I can go to the Net 8 client on the SQL Server and test the connections. All the results are a success.
It looks like some kind of caching is happening when opening a connection using a linked server, Is there a way to force the connection to re-establish?
Thanks,
Craig
March 22, 2006 at 12:19 pm
I always hated those 7399 errors--in BOL they refer to Access databases which is not the case using an Oracle driver 🙂 Anyway, we got the same error on several servers linking to Oracle databases and the fix was to add a startup parameter -g in SQL Server. It seems that the provider runs out of memory and returns that error. Take a look at "Using Startup Options" in BOL (you'll need to stop & start SQL Server). Hope this helps.
Viewing 2 posts - 1 through 1 (of 1 total)
You must be logged in to reply to this topic. Login to reply