Re: Replication sucks :-)

["Pierangelo Masarati" <ando@sys-net.it>]

> Hi !

Welcome to this list.

>
> I tried to setup a simple replication without ssl.
>
> Everything seems to be o.k - but it simply doesn´t work :-)
>
> I see successful replication in the replication-log

this is by no means index of success;
(partial) success is when replica logs
are wiped by slurpd from slapd's replica
log into slurpd's replica logs.

> - but
> the entries on the salve remain in the old state:
> --
> replica: 10.1.1.21:389
> time: 1074242734
> dn: uid=test,ou=EDV,ou=People,dc=testdomain,dc=de
> changetype: modify
> delete: street
> street: werwer
> -
> add: street
> street: Finkenweg 123
> -
> replace: entryCSN
> entryCSN: 2004011608:45:34Z#0x0001#0#0000
> -
> replace: modifiersName
> modifiersName: cn=admin,dc=testdomain,dc=de
> -
> replace: modifyTimestamp
> modifyTimestamp: 20040116084534Z
> -
>
>
> ==> /var/lib/ldap/replog: file truncated <==
> --
>
> There are no real negative entries in the syslog of both systems.
>
> I use Debian/Sarge with OpenLDAP/slapd : 2.1.23-1
>
> I did the following - whats wrong ?
> --
>  master-Server :  10.1.1.11
>  slave-Server  :  10.1.1.21
>
>
>   [master : step 1]
>
>    $$ /etc/init.d/slapd stop
>
>   [Slave  : step 2]
>
>    $$ /etc/init.d/slapd stop
>
>    "/etc/ldap/slapd.conf":
>    --
>    updatedn="cn=admin,dc=testdomain,dc=de"
>    updateref="10.1.1.11:389"

^^^ this is not a valid updateref; RTMF.

>    --
>
>   [master : step 3]
>
>   * copy the database
>
>     $$ slapcat > ~/out
>     $$ scp ~/out root@10.1.1.21:
>
>   * change the config
>
>     "/etc/ldap/slapd.conf":
>
>     global-section:
>     --
>     replogfile      /var/lib/ldap/replog
>     --
>
>     database-definition:
>     --
>     replica uri=ldap://10.1.1.21:389
>             "binddn=cn=admin,dc=testdomain,dc=de"
>              bindmethod=simple credentials=myadminpasswd
>     --
>
>   [slave : step 4]
>
>   $$ cd /var/lib/
>   $$ tar zcvf ../ldap-backup.tgz ldap
>   $$ rm -f ldap/*
>   $$ slapadd -l /root/out -f /etc/ldap/slapd.conf

did you create the directory where the db will reside?
is it /var/lib/ldap or var/lib/ldap/<something>?

>
>   $$ /etc/init.d/slapd start

is it really running, or are you just confident
because the startup script didn't barf?

>
>   [master : step 5]
>   $$ /etc/init.d/slapd start
>   $$ tail -f /var/lib/ldap/replog

Is slurpd running?  Do you actually see changes
getting copied into slurpd's replog file (see
somewhere the replica/ dir, depending on your
configuration)?  Do you see the slave server
being contacted by slurpd?  Would raise your log
level and tell us what you see?  Would you mind
changing the subject of your next email?

p.

-- 
Pierangelo Masarati
mailto:pierangelo.masarati@sys-net.it