Univention Bugzilla – Bug 35261
Don't leave object in listener cache if a listener module fails
Last modified: 2015-01-06 08:46:47 CET
If a join script failes (on delete) the object is left in the cache. If an object with the same DN is added afterwards the cache entry is found (old entryUUID) and the object is then considered as "not existent" on master ldap.
(In reply to Janis Meybohm from comment #0)
> If a [-join script- -> +listener module+] failes ...
After a removal by UUID a second run using the DN is performed.
The cache entry is now also deleted even when a listener module fails.
r51538 | Bug #35261 Listener: Retry update after delete by UUID [YAML]
r51537 | Bug #35261 Listener: Retry update after delete by UUID
PS: As previously the unit tests in doc.34355/ do run find if the S4C is disbaled!
Code: OK (Although I don't really like the new goto handling in change.c)
ucs-test: OK: http://jenkins.knut.univention.de:8080/view/Autotest/job/UCS%203.2-2%20Autotest%20MultiEnv/69/
I was able to reproduce it. I put a raise in the nss.py listener module and created and deleted group objects. The problem is fixed with the new version.