On Mon, 2005-10-31 at 10:21 -0800, Quanah Gibson-Mount wrote:
--On Monday, October 31, 2005 9:51 AM -0800 George Farris
<farrisg@mala.bc.ca> wrote:
> There was a thread about loosing indexes earlier this year. We are
> experiencing the same problem that every once in a while the ldbm
> database must be re-indexed by stopping slapd, running slapindex on the
> database and then starting slapd again. Was there ever a resolution to
> this problem?
>
> We are having to doing this about once per day on Ubuntu Breezy,
> OpenLDAP 2.2.26-3
>
> What would be the reason the ldap server looses indexes? We have not
> modified the slapd.conf file in a long time, have not added or deleted
> any indexes.
>
> Any tips greatly appreciated.
You are using ldbm, and then wondering why you are having problems? :)
Seems somewhat explanatory to me... I suggest upgrading to one of the
more stable and robust backends, like hdb (OpenLDAP 2.3.11 or later if
you want really good performance from it) or bdb (any of OL 2.2 or 2.3).
Well ldbm worked well for me for over a year with Fedora Core 2, however
ok, I can do that. Can you suggest the best way to migrate all the
users information, passwords etc over to a new database? I believe
slapcat doesn't dump all the information, something about timestamps.
It would be great to have everyone show up the next morning and just
login without ever knowing what took place.
I know I can run slapcat to dump to ldif and then reload but maybe there
is a better way????