Re: slapd crashes when adding entries (ITS#2704)

[dieter@dkluenter.de]

Am Sam, 2003-09-13 um 10.46 schrieb Howard Chu:
> Your backtrace shows the stack for an idle thread, not the thread that actually
> crashed, so there is no useful information here.
> 
> At the moment my snapshot of HEAD passes "make test" successfully for back-bdb,
> back-hdb, and back-ldbm. Can you try again with a fresh build?


I have updated HEAD tonight, but the error still occurs.
test002-populate goes into infinite waiting state when populating the
database with ldapadd.
The last lines of master.log

<= ldap_dn2bv(cn=manager,o=university of michigan,c=us,272)=0
<<< dnPrettyNormal: <cn=Manager,o=University of Michigan,c=US>,
<cn=manager,o=un
do_bind: version=3 dn="cn=Manager,o=University of Michigan,c=US"
method=128
conn=1 op=0 BIND dn="cn=Manager,o=University of Michigan,c=US"
method=128
do_bind: Bind preoperation plugin returned 1.
>>> dnPrettyNormal: <(null)>
<<< dnPrettyNormal: <(null)>, <(null)>
conn=1 op=0 BIND dn="(null)" mech=simple (SLAPI) ssf=0

only ctrl c stops the process

connection_closing: readying conn=1 sd=18 for close
connection_close: conn=1 sd=18
conn=1 fd=18 closed
slapd shutdown: waiting for 0 threads to terminate
lt-slapd shutdown: initiated
====> bdb_cache_release_all
lt-slapd shutdown: freeing system resources.
slapd stopped.

I can slapdd entries but not adding or modifying a running system with
ldapmodify or ldapadd.

I notices a strange behaviour with back-meta as well, that is, I only
can do an anonymous search, as soon as I try to bind by means of simple
bind, either as rootdn or user, slapd goes into inifite waiting state as
with slapadd. A strong bind is not possible at all, but with 2.1.19 and
I believe my rewriting rules are correct, at least the work with 2.1.19
But I have to dig deaper into this matter to file an ITS.

-Dieter


-- 
Dieter Kluenter  | Systemberatung
Tel:040.64861967 | Fax: 040.64891521
mailto: dkluenter(at)dkluenter.de
http://www.avci.de