[Date Prev][Date Next] [Chronological] [Thread] [Top]

ghost entry



Hi, I've got:

Solaris 8 (all patches)
openldap-2.1.22
db-4.1.25.NC

Working with a great performance and functionality, the problem is a corrupt
entry, lets say:

dn: cn=badone,o=mycorp

When I search for it on the ldap tree it does not exists, besides, when I
try to add it (caring syntaxes on everything) I get this message saying it
already exists:

Apr  1 17:07:48 siboney slapd[137]: [ID 513 local4.debug] conn=150961 op=2
ADD dn="cn=mavj716j,o=sat"
Apr  1 17:07:48 siboney slapd[137]: [ID 775 local4.debug] => bdb_dn2id_add:
put failed: DB_KEYEXIST: Key/data pair already exists -30997
Apr  1 17:07:48 siboney slapd[137]: [ID 217 local4.debug] conn=150961 op=2
RESULT tag=105 err=68 text=

When I try to delete it (It doesn't exists!):

Apr  1 19:13:41 solaris2 slapd[137]: [ID 307130 local4.debug] conn=1 op=1
DEL dn="cn=mavj716j,o=sat"
Apr  1 19:13:41 solaris2 slapd[137]: [ID 217296 local4.debug] conn=1 op=1
RESULT tag=107 err=32 text=

I've cleaned out the indexes, recreated them and restarted the server.
Actually I had even ran BerkeleyDB utilities to verify and check consistency
of the data bases, they say there's nothing wrong with them (a fact, since
slapd starts and keeps working with any problem).

I've moved on the entire OpenLDAP DB to a testing environment and the
problem still goes on.

The workaround for it was to export the entire DB to LDIF, reinitialize de
DB and importing back again, but this is totally not nice under any
situation  :/

I wish you could tell me if this is a normal behavior with busy servers or
something that could light my life, please.

Best Regards.
HM