2.4.13 slapacl strange behavior

[Dmitriy Kirhlarov <dimma@higis.ru>]

Hi, list.

We are upgrade our openldap installation to 2.4.13 with Berkley DB 47.
We use mirror mode on two openldap servers.

Now we have strange behavior of slapacl (it's look like ITS#3622 issue).

$ sudo slapacl -b o=vega -D cn=ldapadm,o=vega
always return correct data.

$ sudo slapacl -b o=vega -D uid=dkirhlarov,ou=users,o=vega
authcDN: "uid=dkirhlarov,ou=users,o=vega"
entry: write(=wrscxd)
children: write(=wrscxd)
objectClass=organization: write(=wrscxd)
o=vega: write(=wrscxd)
structuralObjectClass=organization: write(=wrscxd)
entryUUID=0e0b8986-9cc3-102b-96f3-3bcca1c2be14: write(=wrscxd)
creatorsName=cn=ldapadm,o=vega: write(=wrscxd)
createTimestamp=20070522151544Z: write(=wrscxd)
description=Vega Enterprise: write(=wrscxd)
entryCSN=20081118152239.589172Z#000000#001#000000: write(=wrscxd)
modifiersName=uid=dkirhlarov,ou=users,o=vega: write(=wrscxd)
modifyTimestamp=20081118152239Z: write(=wrscxd)
contextCSN=20080919162319.000000Z#000000#000#000000: write(=wrscxd)
contextCSN=20090203122647.662600Z#000000#002#000000: write(=wrscxd)
contextCSN=20090203123435.986113Z#000000#001#000000: write(=wrscxd)
bdb(o=vega): Error: closing the transaction region with active transactions
bdb_db_close: database "o=vega": close failed: Invalid argument (22)

after that, database o=vega corrupted:

$ sudo slapacl -b o=vega -D cn=ldapadm,o=vega
hdb_db_open: database "o=vega": unclean shutdown detected; attempting 
recovery.
hdb_db_open: database "o=vega": recovery skipped in read-only mode. Run 
manual recovery if errors are encountered.
hdb_db_open: database "o=vega": alock_recover failed
bdb_db_close: database "o=vega": alock_close failed
backend_startup_one: bi_db_open failed! (-1)
slap_startup failed

slapd restart fix a problem.

Could somebody, please, reproduce this issue and comment it?

Our system:
$ uname -rs; pkg_info -Ix openldap-serv
FreeBSD 7.1-amd64-20090114-RELENG_7_1
openldap-server-2.4.13 Open source LDAP server implementation

WBR.
Dmitriy