Hello,
I am having a small problem with an OpenLDAP master at work. Twice today, at seemingly random intervals, it has stopped responding to all queries, and all clients lose connection to the master, immediately after printing the following message in the debug log:
Jan 7 10:04:58 ldap slapd[20310]: send_search_entry: conn 1851 ber write failed.
There are no more messages printed before or after this for hours.
I looked at the source, but it didn't seem particularly illuminating (it looks like this happens when send_ldap_ber() doesn't write anything and returns 0 bytes, but I don't know how to debug that, since there are a lot of conditions there that return 0).
I saw this[0] thread, but we are already using 2.4.23 on Debian Squeeze, so it appears to not be the same problem:
@(#) $OpenLDAP: slapd 2.4.23 (Dec 16 2012 11:48:44) $ root@carillon:/tmp/buildd/openldap-2.4.23/debian/build/servers/slapd
I am not aware of performing any changes to our LDAP configuration recently.
How should I go about debugging this? Is this a known issue? I searched openldap-bugs, but I didn't find anything recent that seemed related.
Thanks.
0: http://www.openldap.org/lists/openldap-technical/200907/msg00199.html