On 11-09-22 7:37 PM, Howard Chu wrote:

Daniel Qian wrote:

...

The problem direction has the same syncrepl configuration as the working one except for the rid and provider name:

olcSyncrepl: {0} rid=102 provider="ldap://server2.prod:389/" type=refreshAndPersist retry="60 30 300 +" keepalive=1200:10:3 searchbase="dc=mydomain,dc=com" bindmethod=simple binddn="cn=replica,dc=mydomain,dc=com" credentials=xxxxxx starttls=critical tls_cacert="/etc/pki/CA/cacert.pem"

On the consumer side I am seeing these messages:

Sep 22 22:02:02 ldaprov1 slapd[15466]: do_syncrep2: rid=102 got search entry without Sync State control Sep 22 22:02:02 ldaprov1 slapd[15466]: do_syncrepl: rid=102 rc -1 retrying (29 retries left)

and on the provider side I am seeing these:

...

The sync connection is supposed to be persistent but it keeps closing down and reconnecting.

Anyone know what could be the reason?

Obviously your "provider side" doesn't actually have the syncprov overlay configured.

I put the exact same configuration for syncprov on two servers

[root@localhost cn=config]# ldapsearch -LLL -b cn=config -D cn=admin,cn=config -W -ZZ -h ldaprov1.prod cn=module{0} Enter LDAP Password: dn: cn=module{0},cn=config objectClass: olcModuleList cn: module{0} olcModulePath: /usr/lib64/openldap olcModuleLoad: {0}syncprov

[root@localhost cn=config]# ldapsearch -LLL -b cn=config -D cn=admin,cn=config -W -ZZ -h ldaprov2.prod cn=module{0} Enter LDAP Password: dn: cn=module{0},cn=config objectClass: olcModuleList cn: module{0} olcModulePath: /usr/lib64/openldap olcModuleLoad: {0}syncprov

[root@localhost cn=config]# ls /usr/lib64/openldap accesslog-2.4.so.2 constraint-2.4.so.2 dyngroup-2.4.so.2 pcache-2.4.so.2 retcode-2.4.so.2 smbk5pwd-2.4.so.2 translucent-2.4.so.2 accesslog-2.4.so.2.6.0 constraint-2.4.so.2.6.0 dyngroup-2.4.so.2.6.0 pcache-2.4.so.2.6.0 retcode-2.4.so.2.6.0 smbk5pwd-2.4.so.2.6.0 translucent-2.4.so.2.6.0 accesslog.la constraint.la dyngroup.la pcache.la retcode.la smbk5pwd.la translucent.la auditlog-2.4.so.2 dds-2.4.so.2 dynlist-2.4.so.2 ppolicy-2.4.so.2 rwm-2.4.so.2 sssvlv-2.4.so.2 unique-2.4.so.2 auditlog-2.4.so.2.6.0 dds-2.4.so.2.6.0 dynlist-2.4.so.2.6.0 ppolicy-2.4.so.2.6.0 rwm-2.4.so.2.6.0 sssvlv-2.4.so.2.6.0 unique-2.4.so.2.6.0 auditlog.la dds.la dynlist.la ppolicy.la rwm.la sssvlv.la unique.la collect-2.4.so.2 deref-2.4.so.2 memberof-2.4.so.2 refint-2.4.so.2 seqmod-2.4.so.2 syncprov-2.4.so.2 valsort-2.4.so.2 collect-2.4.so.2.6.0 deref-2.4.so.2.6.0 memberof-2.4.so.2.6.0 refint-2.4.so.2.6.0 seqmod-2.4.so.2.6.0 syncprov-2.4.so.2.6.0 valsort-2.4.so.2.6.0 collect.la deref.la memberof.la refint.la seqmod.la syncprov.la valsort.la

Can you tell what is missing on one of the two servers? Is there a command to tell if the module is actually loaded into the execution space?

Thanks, Daniel

On 11-09-22 9:37 PM, Brett @Google wrote:

you need to add the syncrepl overlay to the backend containing your replication source data, to make it syncrepl aware.

On Fri, Sep 23, 2011 at 10:30 AM, Daniel Qian <daniel@up247solution.com mailto:daniel@up247solution.com> wrote:

On 11-09-22 7:37 PM, Howard Chu wrote:

    Daniel Qian wrote:

        The problem direction has the same syncrepl configuration
        as the working
        one except for the rid and provider name:

I am not sure what you mean by "add the syncrepl to the backend". The two nodes have pretty much the same configuration for the backend containing the replicated data but the question is why one is working and the other one is not.

[root@localhost cn=config]# ldapsearch -LLL -b cn=config -D cn=admin,cn=config -W -ZZ -h ldaprov2.prod olcDatabase={1}bdb > /root/ldaprov2.ldif Enter LDAP Password:

[root@localhost cn=config]# ldapsearch -LLL -b cn=config -D cn=admin,cn=config -W -ZZ -h ldaprov1.prod olcDatabase={1}bdb > /root/ldaprov1.ldif Enter LDAP Password:

[root@localhost cn=config]# !diff diff /root/ldaprov1.ldif /root/ldaprov2.ldif 18c18 < olcRootPW: {SSHA}/2ht...................... ---

olcRootPW: {SSHA}lfxH.....................

20c20 < olcSyncrepl: {0}rid=102 provider="ldap://ldaprov2.prod:389/" type=refreshAndP ---

olcSyncrepl: {0}rid=202 provider="ldap://ldaprov1.prod:389/"

type=refreshAndP

above lines are the only difference of the backend configurations in the two nodes.

Thanks, Daniel