Some users reported that their ACLS tabl contains duplicated entries after a while, which could produce too long ACLS and make the READ ACLS cache fail to be
(Currently we don't have a reproduction case.)
A check should be added in org.nuxeo.ecm.core.api.security.ACL#setACEs to emit a WARN when there are duplicate ACPs.
There is now a WARN log when ACL.setACEs contains duplicate entries.
WARN [ACLImpl] Setting an ACL with at least one duplicate entry: ACE(...) , ACL entries: ...
By setting the log4j level to TRACE there is a stack trace available.
<logger name="org.nuxeo.ecm.core.api.security.impl"> <level value="TRACE" /> </logger>