https://bugs.openldap.org/show_bug.cgi?id=9858
Issue ID: 9858 Summary: cn=config replication refresh breaks mdb database Product: OpenLDAP Version: 2.6.2 Hardware: All OS: All Status: UNCONFIRMED Keywords: needs_review Severity: normal Priority: --- Component: slapd Assignee: bugs@openldap.org Reporter: quanah@openldap.org Target Milestone: ---
working in an environment with the following layout:
Config node -> This node contains the cn=config databases for consumers in one tree and providers in another tree.
provider nodes -> Pull their configuration from the config node via standard syncrepl
consumer nodes -> Pull their configuration from the config node via standard syncrepl
This worked fine with OpenLDAP 2.4.
With 2.6, using the same cn=config configuration, slapd will segfault if the cn=config db is put through a REFRESH sequence. I've ruled out syncprov as an issue as I removed the syncprov configuration from the consumer nodes entirely and it still occurs (I was thinking perhaps it was the syncprov checkpoint at shutdown). At this point I can reproduce this trivially 100% of the time simply by having a node REFRESH cn=config.
Once the REFRESH has occurred, slapd will continue running. However the next time it is stopped/started it will SEGV on startup. It is usually possible to use slapcat to extract the MDB database and rebuild it, at which point slapd will start again. A couple of times I was only able to extract a portion of the database using slapcat (70MB of 650MB).