Dear OpenLDAP list users,
I'm having syncrepl oddities. One of my consumer servers has an entry which is absent on the provider. I thought syncrepl would delete it during "present phase" mentioned in the admin guide but it's been sitting there for days and hasn't been removed. There is only one consumer on which this superfluous entry is present. I'm logging sync level on provider and consumers since I identified the problem and nothing odd appears. I unfortunately did not log sync level at the time the problem appeared so maybe I missed any useful info. The only thing I got in the logs when the problem occured was "deferring operation: too many executing" but I also sometimes get the same error in the logs and nothing bad happens to the data.
I'm reading through the guide trying to understand what's going on and in the syncrepl section (http://www.openldap.org/doc/admin23/syncrepl.html ; yes I'm reading the version 2.3 documentation because I'm using the CentOS 5.5 packaged version : 2.3.43-12.el5) I've come across references to a history store maintained on the provider which is useful for the delete phase used in refreshAndPersist, which is what I am using. I thought this might be a plausible clue for my problem. However, nowhere else do I find any reference to this history store, how it's configured or how it works.
Could the use of RefreshAndPersist explain the superflous entry present on one consumer and should(could) I configure a history store to solve this issue? Or should I switch to RefreshOnly to prevent this kind of problem from occuring again in the future? It's the first time I've come across it but when my directories will really be used they will be absolutely critical and I'd rather not have this kind of problem again if possible.
Thanks for your help.
Best regards,
Oliver