[Date Prev][Date Next] [Chronological] [Thread] [Top]

Indexing? Please HELP



Hi!

For 15 Querys (only single objects, nearly no subtree-search) with about 
50 bytes of data per object via html-php-ldap, it takes 1:30 (one and a 
half) minutes for the html-page to load with slapd producing a 
CPU-load of 90%.

My db isn't large - slapcat produces a ldif-file of 385 kbytes... 
That's not really large, not even for a relatively "small" System like
the one i'm using (P2-266 with 96MB Ram), is it?

Is it faster to read the whole subtree and filter on application-level,
or should read only one entry at a time?

Querys are about 80 % like this
"SRCH Base = SomeBaseDN scope=1 filter =
'(myindexedobjectname=objectname)'"
the rest is 
"SRCH Base = SomeBaseDN scope=1 filter =
'(objectclass=someobjectclass)'"
and
"SRCH Base = SomeBaseDN scope=2 filter =
'(objectclass=someobjectclass)'"

Querying known indexed objects is really fast. I Got 20 entries in 4
secs, including php/net-transfer time... seems to be an index problem
then...

How can I check my indexing? I have

index   objectClass eq 
index default pres,eq,sub
index   cn,somespecialobject,anotherspecialobject pres,eq,sub

in slapd.conf and did "slapindex -f /path/to/slapd.conf" after that...
that shold be enough, or did I miss something?

Thanks for your help

Timo Boettcher