openldap-2.4.35: "failed to start slapd"

[Xin Li]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

On 4/28/13 10:30 AM, Chris Rees wrote:
> On 28 April 2013 18:17, Jerry <jerry at seibercom.net> wrote:
>> On Sun, 28 Apr 2013 16:48:33 +0100 Chris Rees articulated:
>> 
>>> Have you tried getting the older version and installing that?
>>> 
>>> Ports-mgmt/portdowngrade will help you do that.
>>> 
>>> Not a proper solution, but it'll get it working again if
>>> you're desperate.
>> 
>> Actually, the database "log" got corrupted. Even the "db-recover"
>> would not correct it. Luckily, I had a semi-recent backup of the
>> directory. I had to run  "/usr/local/libexec/slapd" via "GDB" to
>> find out what the problem was. I have no idea what caused the
>> corruption though. The only thing that comes to mind is that I
>> DID NOT shut down openldap prior to updating it. That never
>> caused a problem before though.
> 
> I'm really glad you had backups.  I'd be surprised if the in-place 
> upgrade broke things, but it might cause a problem.  I've never had
> an issue with any daemons like that.
> 
> Xin Li, do you know if corruption due to this is possible?  Would
> a @stopdaemon line be appropriate here?

This is weird.  We do have some similar construct for the server
package -- if the package is being removed, the daemon would be stopped.

I haven't seen similar issue in the past, what's the BDB version used
with OpenLDAP?

Cheers,

-----BEGIN PGP SIGNATURE-----

iQEcBAEBCAAGBQJRfgRDAAoJEG80Jeu8UPuz3D0IAIVthtvgUOT2+jhyJ3I1uZXp
7xmvKQXUfHXHvsEmu8eLlkjWGaHjuPgE0Ssvap4nkl5NcEENw9YlIIrm8GDCwsC5
uPphCjLhR+RcX0Z6QOjaoHGR3hj29e1nG/Ei9O4pRxiHnPqJz+sZSbIyzrmaeb1v
BAl7df88vlE45e40WuCm9W6QPbaAPqXi9ZMzk9YN9crdB3YKvGMgZT89liS40zQN
uBNRE2sBXJUPhtqjEVE04o9QdIYYIuP9zNH1BFqup4XMybIxdrKCtYbW5iKfkDtB
6cRG014/YxhZRvgRvriaifP6SLXu6pTMIuuFWeC8MQn5uEz9mrczrB5RaHIgwPU=
=FQom
-----END PGP SIGNATURE-----