Answer originally left in comments by David Mann:
If you are unable to run FREETEMPORARY then I believe your only option is to disconnect/reconnect the session. This will cause a cleanup of any temp blob space used by that session. Had a similar issue here where we couldn't touch the app code, so we ended up tweaking their connection pool to expire and reconnect more frequently.
Oracle Metalink Doc ID 802897.1 "How to Release the Temp LOB Space and Avoid Hitting ORA-1652" covers this, for 11g your option is to invoke FREETEMPORARY or to exit session. They also mention a new event 60025 which is supposed to free temp segments when all the temp LOBs for a session are no longer used, but I didn't see any other detail on that.
The application I maintain does something similar that is database OS independent for around 100 users but in a much simpler fashion. Users are authenticated by the application and authorized from the database. Here are the details:
- a table for users including their email and enabled/disabled status
- a table of all AD groups that have a specific prefix to indicate that they are application groups (eg: APP_ADMINS, APP_USERS)
- a table with group/user links
The next part is synchronizing the information. In our organization if new users are added or permissions for existing users are changed IT does the job in the morning. The application typically has a window of low usage overnight (ie: not 24 hours a day/usage).
On the application web server we have a windows service which syncs the groups and users.
On the database at night I run a scheduled job which truncates the group-user links and recreates them by a direct query to Active Directory using DBMS_LDAP.
For your purposes you would want to write a more sophisticated package that has a database package running three LDAP queries to synchronize groups, users and group users instead of my truncate solution.
I found DBMS_LDAP to be powerful but poorly documented. If the IT department should happen to change the OU that users are in you don't get a lot of useful information from the Oracle end.
Tim's site was very useful with this article and here has the code I reused.
Edit: There are two parts to using an application:
- who are you? (authentication) if you do not supply a name and password listed in Active Directory you are not allowed past the log in page.
- what are you allowed to do? (authorization) Here is where the granularity of permissions has a wide range. We have limited ourselves to a basic set of operations on a portion of a unit of work: Read, Update, Assign other users, Unassign other users. Our permissions are not so granular so they can be worded as "User X has select on this table". They can be better said as "User X
can update this portion of the file if they are assigned".
In part this is driven by the application. It creates a pooled connection to the database as required. Users log onto the application but connect to the database as a shared user who has all the permissions required.
To go into a more detailed example:
User X logs in. They have a username and password that matches what is in Active Directory.
After authentication they are taken to the main page. During that time the application queries the database for what permissions this user has and caches them in memory.
The main page loads each sub work area. Each area asks the cached permissions:
- "Do you have authority to see this area?". If so show the work area.
- " Do you have permission to edit this area?" If so show an edit/create link.
Edit2:
The original poster comments that
We are trying to replicate the user permissions onto application to DB
permissions onto database objects.
This sounds like a square peg/round hole problem to me. Your business logic may be defined as User X needs select on this table but if you ask the user they are working with the application. They need to be able to work with whatever the application is displaying. It is common to group tables and views and custom filters to show what the user needs in the application. This is why some applications find it easier to enforce authorization at the application level.
What is it about your application that requires database permissions to be replicated in Active Directory?
Best Answer
I would hope you have some session pooling going on, or at least one database connection per thread. This is not clear from your description. In a multi-threaded design, I would not close connections in the threads unless they make their own connections (which I would consider a bad design).
I would get the process running with one thread, then add an additional thread. Once you are comfortable you have that multi-threading gives you the results you want, you can scale up.
You could turn on auditing on the Oracle side to see what is really happening.
You didn't say how long the queries take to run. If they take minutes, to run you need to ensure you have keep-alive on the network connections. Otherwise, the network connection may get dropped.
EDIT: Check the resource limits applicable to the userid you are connecting with. These might get your connections closed. You may also be getting killed off due to exceeding resource limits. I would expect the reason for disconnect to be in the SQL Error text for the connection.
Given how long these queries are taking, you may want to investigate optimizing the queries. Unless you have lots of memory and processors on the database server you may find that the queries run faster when run serially. You may be forcing data out of the buffer pool with this many long queries running in parallel.