[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: Multiple Backends and SyncRepl



The solution with one rid per OpenLDAP Slave (with 18 different hdb Backends each) works perfectly as you explained :-)

Thank you for your help!

Regards, Michael

Howard Chu wrote:

The rid's should be unique across a server. The only time where this may be a problem is when you specify a sync cookie on the slapd command line (see slapd(8) -c option) as the cookies are identified by their rid; if you try to override a context's CSN "rid=001,csn=xxx" then that override will occur for all of the contexts with the matching rid.

Michael Eichenberger wrote:

Dear List

I've got a single Master, multiple Slave Setup and I want to use SyncRepl for the the replication of my data.

The Master, aswell as the Slaves have multiple hdb Backends defined (each major SubTree has it's own Backend).

Can the "syncrepl rid=xxx" entries be the same per Slave? Example:

database    hdb
suffix          "ou=configuration,o=stepping-stone,c=ch"
rootdn        "cn=Manager,o=stepping-stone,c=ch"
syncrepl     rid=001

database    hdb
suffix          "ou=administration,o=stepping-stone,c=ch"
rootdn        "cn=Manager,o=stepping-stone,c=ch"
syncrepl     rid=001

This setup seems to be working fine, but I've not done a lot of testing yet ...

Or do they need to be different, not only per Slave, but also per backend?

database    hdb
suffix          "ou=configuration,o=stepping-stone,c=ch"
rootdn        "cn=Manager,o=stepping-stone,c=ch"
syncrepl     rid=001

database    hdb
suffix          "ou=administration,o=stepping-stone,c=ch"
rootdn        "cn=Manager,o=stepping-stone,c=ch"
syncrepl     rid=002

Many thanks in advance for any help to clarify this setup!

Kind regards, Michael




-- visit: http://www.stepping-stone.ch -- e-mail: michael.eichenberger@stepping-stone.ch mobile: +41 76 392 36 23 icq: 238901781