Re: Odd problem with open-ldap

[Adam Tauno Williams <adam@morrison-ind.com>]

> It almost seems like the more I select the longer it takes.. for
> instance if I select 7,000 entries, the delay is about 1 1/2 to 2
> seconds.
> And 5,000 entries has a delay of about 1/2 a second.
> Meanwhile doing 700 entries has no delay and the records actively fly
> past my screen as I watch them vanish on the slapd -d 9 debug console.
> I have the following in my config:
> directory       /usr/local/var/openldap-ldbm

Your using an LDBM backend?  DON'T!  Use back-bdb.

> # Indices to maintain
> index   objectClass,x500uniqueIdentifier,uid,mail,mailAlternateAddress  eq

Seems kind of spartan.

> -rw-------    1 root     root      8589312 Dec 30 09:08 dn2id.dbb
> -rw-------    1 root     root     29884416 Dec 30 09:08 id2entry.dbb
> -rw-------    1 root     root         8192 Dec 30 08:18 mailAlternateAddress.dbb
> -rw-------    1 root     root      1548288 Dec 30 09:08 mail.dbb
> -rw-------    1 root     root         8192 Dec 30 08:41 nextid.dbb
> -rw-------    1 root     root       253952 Dec 30 08:54 objectClass.dbb
> -rw-------    1 root     root      1564672 Dec 30 09:08 uid.dbb
> -rw-------    1 root     root      1679360 Dec 30 09:08 x500UniqueIdentifier.dbb
> What's up with that, or what other information is needed?

If you are using back-bdb, you're missing the DB_CONFIG file.  Your
performance will stink, and I'm surprised slapd doesn't just lock up.