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

Bug in deleting entry in MirrorMode?



Hello,

I'm writing to this list to clarify if I discovered a bug or just did not configure OpenLDAP properly.

I've set up OpenLDAP to run in "MirrorMode":

# [...]
overlay syncprov
syncprov-checkpoint     100 10
syncprov-sessionlog     100
# [...]
syncrepl        rid=001
                provider=ldap://localhost:11389
                type=refreshAndPersist
                searchbase="o=esolution,dc=deutscherv,dc=de"
                schemachecking=on
                bindmethod=simple
                binddn="cn=MirrorMode,o=esolution,dc=deutscherv,dc=de"
                credentials="M1rr0rM3"
                retry="60 +"

serverID        1
mirrormode      on
# [...]

The second server is set up identically, except "serverID" is 2 and "provider" is set to URL of first server. The fact addresses are "localhost" is because I replicated my main setup on my private server to better inspect and debug the problem.

When adding, modifying and deleting an entry it can happend I do get a positive return from delete, but a subsequent "add" fails with "code=68: Already exists".

This does not happen all the time, but often enough to name it "reproducible".

This seems not to happen if there's a "sleep" between modify and delete. It also seems not to happen if secondary server is down. So either I've got a configuration problem or there's a problem with synchronization.

Anybody here to help, tracking down this problem?

Thanks a lot, and
--
Regards,
Peter