reconciliation does not seem to work ok for some CI types.
i see some strange behavior on my 10.22 cup 6 cp 22 (yeah, quite old one). There's an integration we push from one ucmdb to another one, and in the second one i see bunch of duplicate CIs of some types. For example, IPs are ok, nodes are ok, but filesystem and cpu are not. The rules are like OOB, but the filesystems with the same mountpoint, CPUs with the same cpuid and same container are much affected. I even checked the duplicates in the database by the very concrete queries - the duplicates are there! But if i try to add one more filesystem linked to the same root in IT univerce - it fails, saying 'data already exists', so this kind of data-in is ok. I purged the exeption table in JMX just in case, rebooted for sure, but it still tekes place. Going to try 'newdbtool', and created an enrichment to purge the bad stuff, BUT i can't understand what is not ok with the engine! The reconciliation log in standard loglevel and error.log don't capture the errors at that time.
Any ideas? Maybe some known issue for 10.22?
Re: reconciliation does not seem to work ok for some CI types.
We are still on 10.22 CUP6 and doing tonns of syncs between instances almost every day. No OTTB problems like this tracked.
Check that reconciliation rules are in sync between systems, especially for CITs with the problems.
If you planing to sync between two UCMDBs in a frequent manner, I would suggest considering setting Global ID sync between instances. It's available on 10.2X as well.
-Dmitry Gomel, PMP
Click the Like button at the bottom to say 'Thanks'.