h.b.furuseth@usit.uio.no wrote:
I wrote:
Just reproduced it. Out of locks (so of course configure more), then "Accessing a corrupted shared library". Will file a separate ITS, it doesn't sound like slapadd in a clean directory should need 1000+ locks.
...yes it should, it was a case of writing an entry with many values in an indexed attribute - after the index was big enough that presumably few of the new values' hashes ended up on the same DB page. So, it's just a spurious scary error message afterwards from BerkeleyDB.
The "Accessing a corrupted shared library" message occurs on Linux because it's logging an LDAP result code (80, LDAP_OTHER) as a system errno code. Probably ought to change that...