kill -INT corrupts database (ITS#1982)

[quanah@stanford.edu]

Full_Name: Quanah Gibson-Mount
Version: 2.1.3
OS: Solaris 8
URL: ftp://ftp.openldap.org/incoming/
Submission from: (NULL) (171.64.13.58)


Running:
OpenLdap-2.1.3
OpenSSL-0.9.6d
Cyrus-SASL-2.1.5
BerkeleyDB-4.0.14

We have a master/slave configuration.  We notice that on the slave machine, that
after replication occurs, if you stop slapd with a kill -INT, when you go to
restart it, it fails to start due to the database being corrupt.  This scenario
is repeatable.

--Quanah