RE: slapd presence index option working?

["Kurt D. Zeilenga" <Kurt@OpenLDAP.org>]

At 06:35 PM 9/27/99 -0700, Trevor Porter wrote:
>> Trevor Porter wrote:
>> > 
>> > > At 04:18 PM 9/27/99 -0700, Trevor Porter wrote:
>> > > >My database has around 86K entries and it appears to be
>> > > searching the entire
>> > > >directory for the above query. Any ideas?
>> > >
>> > > Likely the attribute type is present on more than the 4000 entries
>> > > (the default SLAPD_LDBM_MIN_MAXIDS), and hence the index is being
>> > > replaced with an ALLIDS index.  It's quite pointless to generate
>> > > such an index.
>> > 
>> > Agreed, however the attribute type is currently present in 
>> just 35 entries
>> > which is why it appears odd (sorry, I should have mentioned 
>> this in my first
>> > message). I can see ldapsearch spit out entries very slowly 
>> rather than
>> > quickly using the contents of the presence index file to 
>> grab the entries.
>> > 
>> > ...Trevor
>> 
>> Did you generate the index using ldbmcat and ldbmindex ?
>
>I've tried both ldif2ldbm and ldif2index. In both cases, ldif input was
>generated cleanly using ldbmcat. I'm pretty sure I'm following the correct
>procedure since I've have no issues with other index types.

ldbmcat, of course, only repeats what's in id2entry.  If ldbm2ldif
was used previously to add garbage to id2entry, ldbmcat won't fix it.

Anways, assuming you have valid LDIF, wipe the database directory
and rebuild using ldif2ldbm.  Then use ldbmtest to verify that the
index was (or wasn't) created correctly.

Kurt