slapd dies when given 'incorrect' requests

["Peter Hicks" <Peter.Hicks@POGGS.CO.UK>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Folks,

I've found a reproducable bug in slapd from OpenLDAP 1.2.1 - when using 
PHP3/Apache to retreive data and publish it to a web interface, if I try to 
retreive information 'too fast' (or what seems like it), one of the three 
slapd processes turns in to a zombie, LDAP stops responding, the zombie'd 
process can't be killed by a "kill -9 <pid>", and I have to "killall -9 
slapd" and restart slapd to get things working.

The output from syslog just before the sudden death is as follows:

Apr 13 00:08:29 foxbase slapd[24525]: conn=2 fd=7 connection from 
foxbase.poggs.com (10.180.250.3) accepted. 
Apr 13 00:08:29 foxbase slapd[24538]: conn=2 op=0 BIND dn="" method=128 
Apr 13 00:08:29 foxbase slapd[24538]: conn=2 op=0 RESULT err=0 tag=97 
nentries=0 
Apr 13 00:08:29 foxbase slapd[24539]: conn=2 op=1 SRCH base="O=POGGS COMPUTER 
SERVICES,C=UK" scope=2 filter="(cn=STEPH WALKER)" 
Apr 13 00:08:29 foxbase slapd[24539]: conn=2 op=1 RESULT err=0 tag=101 
nentries=1 
Apr 13 00:08:29 foxbase slapd[24540]: conn=2 op=2 UNBIND 
Apr 13 00:08:29 foxbase slapd[24540]: conn=2 op=2 fd=7 closed errno=0 
Apr 13 00:08:35 foxbase slapd[24525]: conn=3 fd=7 connection from 
foxbase.poggs.com (10.180.250.3) accepted.

Can anyone advise on a suitable troubleshooting route?



Peter.



-----BEGIN PGP SIGNATURE-----
Version: PGP 6.0.2 -- QDPGP 2.60 

iQA/AwUBNxNO/yNeTDpbBy8EEQJfzACg/lBl6R83JtOkQQG0DhP3tZjFoWMAoP8B
s7h7UmYBbvMRjen/IQ3W/rZO
=pVqm
-----END PGP SIGNATURE-----

---
Peter Hicks
Network Engineer
peter <DOT> hicks (AT) poggs.co.uk