syncRepl oddities (ITS#2950)

[quanah@stanford.edu]

Full_Name: Quanah Gibson-Mount
Version: 2.2.5
OS: Solaris 8
URL: ftp://ftp.openldap.org/incoming/
Submission from: (NULL) (171.64.19.82)


The slapd on one of my slave systems suddenly segfaulted and got restarted.

In the logs I see this:

Feb  4 16:48:25 ldap-dev1.Stanford.EDU [534]: [ID 870808 local4.debug]
syncrepl_message_to_entry: mods2entry (suMailDrop: value #0 provided more than
once)

Note that this server had been up about an hour before this error occurred, and
that *no* modifications to the master were going on.  Also note that all of the
servers were loaded from the *same* database via slapadd with *no* complaints
from slapadd.

I would assume this means there is a value of "suMailDrop" in the directory for
an entry that appears twice.. like

cn=quanah,cn=accounts,dc=stanford,dc=edu
sumaildrop: quanah
sumaildrop: quanah

How our database would get into that state puzzles me (given that our master
should reject any modification like that).

The fact that a slave's slapd segfaults on this is also not good.

--Quanah