I've worked with several customer environments that implemented directory naming. I'll try to answer the best that I can.
1. If you have 30 development databases and are always creating new ones, dropping old ones and plan to grow the environment significantly in the future, then I'd consider it. If your environment is relatively static, then I don't know that I would. You're already doing more than most environments I've seen by putting a shared tnsnames.ora on a file server--most sites I've seen haven't gotten that figured out yet (though I try to help them in that direction). 2. No, directory naming has nothing to do with the contents of your AD directory. There was a time when Oracle supporting directory naming using AD as the repository--there's a good reason they no longer support it--it plain didn't work in most of the cases where I tried. There's no reason to sync AD with OID either. The OID data you're considering storing for naming has no benefit to being stored in AD. You'll obviously want to back up your OID data via normal database backup methods, but that's about as concerned as you should need to be. 3. You'll need to create an ldap.ora file in the OH/network/admin (IIRC, its location is NOT overridden by TNS_ADMIN) directory. It contains the hostname and port number(s) of your LDAP server(s) along with the default context used in OID to locate your names.
Sounds like a good blog entry...I'll try to get to it soon. In the meantime, Daniel Liu did a good presentation at OOW a few years ago on it. Google found the whitepaper at http://www.bgaarsoe.com/Beacon/BlogDocuments/7d62619d611ca/1457_wp.pdf
Dan Li Li wrote:
Hi List, Anybody using directory naming here? We have about 30 databases and use local naming and most clients are configured to share 1 tnsnames.ora file on a shared file server. I was tasked to look into directory naming to see if we need to implement it. I have spent some time reading "Oracle(R) Database Net Services Administrator's Guide 10g Release 2 (10.2)" and got a few questions: 1. with around 30 databases, is it worth implementing directory naming? 2. We use Active Directory and majority of our database servers are on Unix/Linux, based on the above document, we cannot use Active Directory, instead we have to use OID. If we do implement directory naming, do we need to synchronize AD and OID? if so, how? 3. In directory naming, you specify NAMES.DIRECTORY_PATH=(ldap) in sqlnet.ora file, how does the client find where the directory server is? Any help or pointing me to the correct document to read would be greatly appreciated! Thanks, -Li -- //www.freelists.org/webpage/oracle-l
-- //www.freelists.org/webpage/oracle-l
