--On Thursday, November 17, 2005 6:31 PM -0800 Moe <[EMAIL PROTECTED]> wrote:
Hi everyone, Im having a problem in my replication. I have 3 servers, one has ldap master and the other two are ldap slaves. The master ldap uses slurpd to replicate to the two slaves. The two slave boxes are almost exactly the same (hardware, same ldap version), except that one of them is in a remote location while the master and first slave ldap are in the same location. my system has beed fine for the past 4 months. recently i noticed that the remote ldap slave has like 17 missing entries that were not replicated. The hostname.rej log does not show me why the entries were rejected. All i see is lot of entries saying: Type or value exists: modify/add: vipFuncarea: value #0 already exists. im struggling to find out why this happened. My understanding is that if replication fails slurpd retry to send the replication but that's didn't happen. I have lots of entries in my database and im finding it hard to track the issue , the first ldap slave has the same entries as the master , no porblems at all
Are you sure the entries are missing, or are they simply already there? I've noticed an issue with slurpd where if the response time from the slave on getting an update is too long, it will try to send the update again, which will fail because the slave already got it. The errors you are reporting in your hostname.rej file tend to match that scenario. I would see if there is really any difference between the two replicas for those entries or not.
--Quanah -- Quanah Gibson-Mount Principal Software Developer ITSS/Shared Services Stanford University GnuPG Public Key: http://www.stanford.edu/~quanah/pgp.html
