Jap, changing the order did not really help. When calling netstat, I see the connection is still ESTABLISHED, even if the second machine is switched off. May be I need to wait for the tcp_keepalive_timeout. I would expect the slapd somehow recognizes the loss of the connection to the second slapd.
-----Ursprüngliche Nachricht----- Von: openldap-technical-bounces+thorsten.mueller=aachen.utimaco.de@OpenLDAP.org [mailto:openldap-technical-bounces+thorsten.mueller=aachen.utimaco.de@OpenLDAP.org] Im Auftrag von Dieter Kluenter Gesendet: Freitag, 26. März 2010 15:40 An: openldap-technical@openldap.org Betreff: Re: AW: syncrepl connection / reconnect
"Dieter Kluenter" dieter@dkluenter.de writes:
Thorsten Mueller Thorsten.Mueller@aachen.utimaco.de writes:
Here you are, the config of the second machine is identical, apart from the different provider
[...]
syncprov-checkpoint 100 10
syncprov-sessionlog 100
# syncrepl directiv
syncrepl rid=001
provider=ldap://192.168.120.237:388
bindmethod=simple
binddn="cn=Replicator,dc=DatabaseReplication,dc=head"
credentials="fdet2zS3"
searchbase="dc=head"
starttls=critical
tls_cacert=/etc/TLS/ca-certs/trusted_CAs.pem
tls_cert=etc/TLS/client/client.pem
tls_key=etc/TLS/client/client_key.pem
schemachecking=on
type=refreshAndPersist
retry="5 12 60 +"
mirrormode on
[...]
order matters, syncrepl has to be declared within a database specification, modules are loaded last, after all database specific configuration parameters.
Sorry my bad, this should read ..overlays are loaded last...
-Dieter