[Kolab-devel] [issue1860] replication openldap -> kolad is not working (on CentOS5)

[Bernhard Reiter]

New submission from Bernhard Reiter <bernhard at intevation.de>:

On CentOS 5 with Kolab Server 2.1.0     
tboernert always sees problems with replication from openldap to kolabd.     
Result: Users are not deleted anymore. Creating of users happen     
when kolabd does a full resync, so it is mainly the deletion that does    
not work.    
    
(This is split out issue from    
  kolab/issue1068 (sometimes mailboxes are not created until kolabd restart)    
    
    
Does somebody else use the combination CentOS 5 + Kolab Server 2.1.0?    
If yes: Do you have the same problems?    
    
Details (copied over from issue1068)   
 There are no firewall rules active and no selinux active.   
kolab 2.0.99 (20060619) works.   
   
/sbin/iptables -L -n   
Chain INPUT (policy ACCEPT)   
target     prot opt source               destination            
   
Chain FORWARD (policy ACCEPT)   
target     prot opt source               destination            
   
Chain OUTPUT (policy ACCEPT)   
target     prot opt source               destination   
   
getenforce   
Disabled   
   
kolabd is running and listening to port 9999.  
there comes no traffic from ldap to port 9999 to kolabd.  
  
og_level to 100 (kolab.globals) 
btw: i've no slaves, only one master. 
now delete an existing user, and .... nothing in syslog. 
after restarting kolabd i've got a log of debug and the user 
is deleted. 
 
i'd do some debugs with slurpd and slapd and ... 
slurpd: copy replog "/kolab/var/openldap/replog" to 
"/kolab/var/openldap/openldap-slurp/replica/slurpd.replog" 
Replica 127.0.0.1:9999, skip repl record for cn=hans 
Dampf,dc=intra,dc=tbits,dc=de (old) 
 
why skip repl record ? that can be the problem .... 
 
i saw also no communication to localhost:9999 (kolabd) with 
tcpdump. 
 
in the slurpd.replog are 2 interesing entries: 
once an changetype add with the create of my user 
and once an changetype modify with the kolabdeleteflag 
of my user. 
 
in the slapd debug file i got some strange errors and 
failed, as 
 
ber_get_next 
do_search 
ber_scanf fmt ({miiiib) ber: 
ldap_read: want=8 error=Resource temporarily unavailable 
ber_dump: buf=0x09121618 ptr=0x0912161b end=0x09121685 len=106 
ber_get_next on fd 20 failed errno=11 (Resource temporarily unavailable) 
 
bdb_idl_fetch_key: [b49d1940] 
<= bdb_index_read: failed (-30989) 
<= bdb_equality_candidates: id=0, first=0, last=0 
<= bdb_filter_candidates: id=0 first=0 last=0 
=> bdb_filter_candidates 
 
i think the problem is the replication to kolabd.

----------
messages: 11281
nosy: bernhard, steffen, tboernert, thomas, wilde
priority: urgent
status: need-eg
title: replication openldap -> kolad is not working (on CentOS5)
topic: kksa, server
________________________________________________
Kolab issue tracker <kolab-issues at intevation.de>
<https://intevation.de/roundup/kolab/issue1860>
________________________________________________