Kolabd crashes after deleting user

[Rouven Sacha]

Hi,

i 've successfully set up kolabd 2.2.0-20080709.dfsg-1 on lenny from the
alioth sanpshot repositories. Kolabd crashes reproducibly when changing
any entry in the kolab webinterface. It seems to be connected to
syncrepl reconnection error, but i can't seem to debug it any further
than that. The crash happens roughly 20s after i delete a user in the
webinterface. Following the log, the user gets deleted from ldap, kolab
successfully syncs with ldap and crashes:

Sep 30 11:37:40 bserver-netsrv kolabconf: T Debug: Changing permissions
of /etc/imapd.group to 0640 
Sep 30 11:37:40 bserver-netsrv kolabconf: T Debug: Changing owner
of /etc/imapd.group to root:root (0:0) 
Sep 30 11:37:40 bserver-netsrv kolabconf: T Debug: Finished building
Cyrus groups 
Sep 30 11:37:40 bserver-netsrv kolabconf: T Debug: Finished regenerating
configuration files 
Sep 30 11:37:40 bserver-netsrv kolabconf: KC: Reloading kolab
components 
Sep 30 11:37:40 bserver-netsrv kolabconf: K: Reload finished 
Sep 30 11:37:40 bserver-netsrv kolabconf: KC: Finished 
Sep 30 11:37:40 bserver-netsrv kolabd: SYNCREPL Debug: Finished
Kolab::LDAP::sync sleeping 1s 
Sep 30 11:37:41 bserver-netsrv kolabd: SYNCREPL Debug: Finished
Net::LDAP::Search::sync sleeping 10s 
Sep 30 11:37:51 bserver-netsrv kolabd: SYNCREPL Debug: Control created 
Sep 30 11:37:51 bserver-netsrv kolabd: SYNCREPL Debug: Search created 
Sep 30 11:37:51 bserver-netsrv slapd[1595]: conn=4038 op=12 SRCH
base="dc=buero20,dc=org" scope=2 deref=2 filter="(objectClass=*)" 
Sep 30 11:37:51 bserver-netsrv slapd[1595]: conn=4038 op=12 SRCH attr=*
entryUUID modifytimestamp cyrus-userquota kolabdeleteflag 
Sep 30 11:37:51 bserver-netsrv slapd[1595]: conn=4038 op=12 SEARCH
RESULT tag=101 err=0 nentries=0 text= 
Sep 30 11:37:51 bserver-netsrv kolabd: SYNCREPL Debug: No control:
mesg->code = `84', mesg->msg = `LDAP decode error' 
Sep 30 11:37:51 bserver-netsrv kolabd: SYNCREPL: Aborting 
Sep 30 11:37:51 bserver-netsrv slapd[1595]: conn=4038 op=13 UNBIND 
Sep 30 11:37:51 bserver-netsrv slapd[1595]: conn=4038 fd=34 closed 
Sep 30 11:37:51 bserver-netsrv kolabd: SYNCREPL Debug: No control:
mesg->code = `88', mesg->msg = `Explicit disconnect' 
Sep 30 11:37:51 bserver-netsrv kolabd: SYNCREPL: Aborting 
Sep 30 11:37:51 bserver-netsrv kolabd: K: Abnormal child exit status
encountered, aborting 
Sep 30 11:37:51 bserver-netsrv kolabd: Kolab is shutting down 
Sep 30 11:37:51 bserver-netsrv kolabd: K: SIGINT/SIGTERM detected,
kill()ing children 
Sep 30 11:37:51 bserver-netsrv kolabd: L: Shutting down 
Sep 30 11:37:51 bserver-netsrv kolabd: K: Exiting 


Openldap is 2.4.10-3 straight from Debian Testing with the following
settings (generated by kolab):

#### Provide the modern syncprov/syncrepl method of ldap replication
# This database is a synchronisation provider
# Note that a database can be both a consumer and a provider 
moduleload      syncprov
overlay syncprov
# Guarantee that contextCSN gets written.
syncprov-checkpoint  1024 16
# Save a log of last write operations
syncprov-sessionlog 4096
# Support delta-based syncrepl
syncprov-reloadhint TRUE
# Put an index on attributes used for synchronisation
# Note that these indexes are used locally both for server and client 
# during synchronisation.
index entryCSN eq
index entryUUID eq


Any hints on that issue or should i send that to the debian kolab
maintainers?

Cheers,

Rouven