Hello, I'm somewhat not experienced with LDAP on the server side of things I’m importing openldap 2.4. into 2.6.3. (rockylinux 9). My goal is to 2 have 2 N-way (or multi-master*) ldap nodes. I’ve changed hdb to mdb, created accesslog folder, fixed permissions, SSL etc The import doesn’t throw any errors. My understanding is that I need to have cn=config replication, as well as my small dc=domain,dc=com, replication as well The cn=config replication I call via this on both nodes followed by restarts
dn: cn=config changetype: modify replace: olcServerID olcServerID: 1
dn: olcOverlay=syncprov,olcDatabase={0}config,cn=config changetype: add objectClass: olcOverlayConfig objectClass: olcSyncProvConfig olcOverlay: syncprov olcSpNoPresent: TRUE olcSpReloadHint: TRUE
dn: olcDatabase={0}config,cn=config changetype: modify replace: olcSyncRepl olcSyncRepl: rid=002 provider=ldaps://prod-ldap2.domain.com:636 bindmethod=simple binddn="cn=admin,dc=domain,dc=com" credentials=N… searchbase="cn=config" schemachecking=on type=refreshAndPersist retry="10 10 60 +" tls_reqcert=allow keepalive=240:10:30 olcSyncRepl: rid=001 provider=ldaps://prod-ldap1.domain.com:636 bindmethod=simple binddn="cn=admin,dc=domain,dc=com" credentials=N…… searchbase="cn=config" schemachecking=on type=refreshAndPersist retry="10 10 60 +" tls_reqcert=allow keepalive=240:10:30 - add: olcMirrorMode olcMirrorMode: TRUE
Now once I do that I’ve experimented with changing the olcLogLevel and it seems to work. The rid’s on each node are different server2 has rid=002, server 1 has rid=001 as well as different olcServerID
The part I run into issues is when I enable replication to the dc=domain,dc=com via
dn: olcOverlay=syncprov,olcDatabase={2}mdb,cn=config changetype: add objectClass: olcOverlayConfig objectClass: olcSyncProvConfig olcOverlay: {2}syncprov olcSpCheckpoint: 20 10 olcSpSessionlog: 10000000
dn: olcDatabase={1}mdb,cn=config changetype: modify replace: olcSyncRepl olcSyncRepl: rid=021 provider=ldaps://prod-ldap1.domain.com:636 bindmethod=simple binddn="cn=admin,dc=domain,dc=com" credentials=N…. searchbase="dc=domain,dc=com" logfilter="(&(objectClass=auditWriteObject)(reqResult=0))" schemachecking=on type=refreshAndPersist retry="5 10 60 +" tls_reqcert=allow keepalive=240:10:30 olcSyncRepl: rid=022 provider=ldaps://prod-ldap2.domain.com:636 bindmethod=simple binddn="cn=admin,dc=domain,dc=com" credentials=N…. searchbase="dc=domain,dc=com" logfilter="(&(objectClass=auditWriteObject)(reqResult=0))" schemachecking=on type=refreshAndPersist retry="5 10 60 +" tls_reqcert=allow keepalive=240:10:30 - add: olcMirrorMode olcMirrorMode: TRUE
I have 2 sets rids 001/002 and 021/022 and I have olcMirrorMode set to true on both cn=config and domain replication
I’m pasting the relevant code around accesslog and syncprov that I think I’m getting wrong
dn: olcOverlay={3}syncprov,olcDatabase={1}mdb,cn=config objectClass: olcOverlayConfig objectClass: olcSyncProvConfig olcOverlay: {3}syncprov olcSpNoPresent: TRUE olcSpReloadHint: TRUE structuralObjectClass: olcSyncProvConfig
dn: olcOverlay={4}accesslog,olcDatabase={1}mdb,cn=config objectClass: olcOverlayConfig objectClass: olcAccessLogConfig olcOverlay: {4}accesslog olcAccessLogDB: cn=accesslog olcAccessLogOps: writes olcAccessLogPurge: 07+00:00 01+00:00 olcAccessLogSuccess: TRUE structuralObjectClass: olcAccessLogConfig
dn: olcDatabase={2}mdb,cn=config objectClass: olcDatabaseConfig objectClass: olcMdbConfig olcDatabase: {2}mdb olcDbDirectory: /var/lib/ldap/accesslog olcSuffix: cn=accesslog olcRootDN: cn=admin,dc=adaptavist,dc=com olcDbIndex: default eq olcDbIndex: entryCSN,objectClass,reqEnd,reqResult,reqStart structuralObjectClass: olcMdbConfig
dn: olcOverlay={0}syncprov,olcDatabase={2}mdb,cn=config objectClass: olcOverlayConfig objectClass: olcSyncProvConfig olcOverlay: {0}syncprov olcSpNoPresent: TRUE olcSpReloadHint: TRUE structuralObjectClass: olcSyncProvConfig
Replication works from node1 to node2, and in reverse. But it stops after 20 minutes or so After replication stops I see the on accesslog on one node has 4 records, on the other it has 3 and it never catches up even if I restart although at first it all works regardless of which node I update (change a random password)
What am I doing wrong? Perhaps more than one thing
Thank you
--On Wednesday, January 24, 2024 8:28 AM +0200 Viktor Keremedchiev vkeremedchiev@adaptavist.com wrote:
Hello, I'm somewhat not experienced with LDAP on the server side of things I'm importing openldap 2.4. into 2.6.3. (rockylinux 9). My goal is to 2 have 2 N-way (or multi-master*) ldap nodes. I've changed hdb to mdb, created accesslog folder, fixed permissions, SSL etc The import doesn't throw any errors. My understanding is that I need to have cn=config replication, as well as my small dc=domain,dc=com, replication as well
It is not required to have cn=config replication. And I would note that OpenLDAP 2.6.3 is fairly old at this point with significant fixes done to the 2.6 series since its release. I'd advise using a current release of OpenLDAP 2.6.
The cn=config replication I call via this on both nodes followed by restarts
dn: cn=config changetype: modify replace: olcServerID olcServerID: 1
Each server must have its own, unique, serverID. If you are going to use cn=config replication, then you *must* use the
olcServerID: # URI
format.
Now once I do that I've experimented with changing the olcLogLevel and it seems to work. The rid's on each node are different server2 has rid=002, server 1 has rid=001 as well as different olcServerID
RIDs must be unique INSIDE a particular server, but different servers can use the same RID values.
What am I doing wrong? Perhaps more than one thing
I'd suggest starting with just getting back-mdb replication working between the nodes.
Side note, your configuration for the accesslog DB is missing an index on 'reqDN'.
--Quanah
We stay in the 2.5 LTS branch. 2.6 is more of a Dev Branch if I understand it correctly.
From: Quanah Gibson-Mount quanah@fast-mail.org Sent: Wednesday, January 24, 2024 11:39 AM To: Viktor Keremedchiev vkeremedchiev@adaptavist.com; openldap-technical@openldap.org Subject: [EXTERNAL] Re: Openldap 2.4 -> Openldap 2.6.3 replication hurdles
--On Wednesday, January 24, 2024 8: 28 AM +0200 Viktor Keremedchiev <vkeremedchiev@ adaptavist. com> wrote: > Hello, > I'm somewhat not experienced with LDAP on the server side of things > I'm importing openldap 2. 4. into 2. 6. 3.
--On Wednesday, January 24, 2024 8:28 AM +0200 Viktor Keremedchiev
<vkeremedchiev@adaptavist.commailto:vkeremedchiev@adaptavist.com> wrote:
Hello,
I'm somewhat not experienced with LDAP on the server side of things
I'm importing openldap 2.4. into 2.6.3. (rockylinux 9). My goal is to 2
have 2 N-way (or multi-master*) ldap nodes. I've changed hdb to mdb,
created accesslog folder, fixed permissions, SSL etc
The import doesn't throw any errors. My understanding is that I need to
have cn=config replication, as well as my small dc=domain,dc=com,
replication as well
It is not required to have cn=config replication. And I would note that
OpenLDAP 2.6.3 is fairly old at this point with significant fixes done to
the 2.6 series since its release. I'd advise using a current release of
OpenLDAP 2.6.
The cn=config replication I call via this on both nodes followed by
restarts
dn: cn=config
changetype: modify
replace: olcServerID
olcServerID: 1
Each server must have its own, unique, serverID. If you are going to use
cn=config replication, then you *must* use the
olcServerID: # URI
format.
Now once I do that I've experimented with changing the olcLogLevel and
it seems to work. The rid's on each node are different server2 has
rid=002, server 1 has rid=001 as well as different olcServerID
RIDs must be unique INSIDE a particular server, but different servers can
use the same RID values.
What am I doing wrong? Perhaps more than one thing
I'd suggest starting with just getting back-mdb replication working between
the nodes.
Side note, your configuration for the accesslog DB is missing an index on
'reqDN'.
--Quanah
--On Wednesday, January 24, 2024 4:42 PM +0000 Bradley T Gill bgill@aep.com wrote:
We stay in the 2.5 LTS branch. 2.6 is more of a Dev Branch if I understand it correctly.
2.5 is the current LTS release. 2.6 was the new 'feature' branch. It may become the next LTS. I've personally run it in production at a high volume 100% uptime environment for over 2 years.
--Quanah
On 24 Jan 2024, at 18:39, Quanah Gibson-Mount quanah@fast-mail.org wrote:
--On Wednesday, January 24, 2024 8:28 AM +0200 Viktor Keremedchiev vkeremedchiev@adaptavist.com wrote:
Hello, I'm somewhat not experienced with LDAP on the server side of things I'm importing openldap 2.4. into 2.6.3. (rockylinux 9). My goal is to 2 have 2 N-way (or multi-master*) ldap nodes. I've changed hdb to mdb, created accesslog folder, fixed permissions, SSL etc The import doesn't throw any errors. My understanding is that I need to have cn=config replication, as well as my small dc=domain,dc=com, replication as well
It is not required to have cn=config replication. And I would note that OpenLDAP 2.6.3 is fairly old at this point with significant fixes done to the 2.6 series since its release. I'd advise using a current release of OpenLDAP 2.6.
Got it
The cn=config replication I call via this on both nodes followed by restarts
dn: cn=config changetype: modify replace: olcServerID olcServerID: 1
Each server must have its own, unique, serverID. If you are going to use cn=config replication, then you *must* use the
olcServerID: # URI
If I use olcServerID: 1 ldaps://prod-ldap1.domain.com - server doesn’t start post ldapmodify.
format.
Now once I do that I've experimented with changing the olcLogLevel and it seems to work. The rid's on each node are different server2 has rid=002, server 1 has rid=001 as well as different olcServerID
RIDs must be unique INSIDE a particular server, but different servers can use the same RID values.
Got it
What am I doing wrong? Perhaps more than one thing
I'd suggest starting with just getting back-mdb replication working between the nodes.
Side note, your configuration for the accesslog DB is missing an index on 'reqDN'.
Added
--Quanah
I think I’ve been trying to add syncprov and acceslogs overlays that I already have dn: olcOverlay={3}syncprov,olcDatabase={1}mdb,cn=config dn: olcOverlay={4}accesslog,olcDatabase={1}mdb,cn=config
but then again adding unnecessarily dn: olcOverlay=syncprov,olcDatabase={2}mdb,cn=config
I’ve also spotted in the logs that if I have olcMirrorMode twice for cn=config and the domain - the log says it is already enabled. But if I remove second one it is telling me that I’m missing referral when ldapsearch/update any of the nodes.
Viktor
--On Thursday, January 25, 2024 9:48 AM +0200 Viktor Keremedchiev vkeremedchiev@adaptavist.com wrote:
If I use olcServerID: 1 ldaps://prod-ldap1.domain.com - server doesn't start post ldapmodify.
Sounds like this doesn't match the options passed to slapd at startup. I do note your missing a port at the end of the URI which may be why.
--Quanah
openldap-technical@openldap.org
-
Bradley T Gill -
Quanah Gibson-Mount -
Viktor Keremedchiev