It�s something estrange
because when it�s only doing query
./ldasearch -b "o=ggc" 'bid=100000*'
if i do another query if runs perfectly (bid is a non standar attribute with
all indexes).
and i have probed it before replication ant it happens the same thing.
I try to reimport many times the ldif, but the query persists very slow (and
only there are 4 entries with that filter).
I read about nsslapd-allidsthresold and my inicial value was 4000.
I increased it to 16000 and then reimport and it seems i have do nothing,
because in log i see notes=U
This is the acces log for the consult:
[05/Dec/2002:18:57:14 +0100] conn=9 op=0 SRCH base="o=ggc" scope=2
filter="(bid=10000*)" attrs=ALL
[05/Dec/2002:19:01:38 +0100] conn=9 op=1 UNBIND
[05/Dec/2002:19:01:38 +0100] conn=9 op=1 fd=27 closed - U1
[05/Dec/2002:19:01:38 +0100] conn=9 op=0 RESULT err=0 tag=101
nentries=4 etime=264 notes=U
And the notes=U appears too.
"Rich Megginson" <[EMAIL PROTECTED]> escribi� en el mensaje
[EMAIL PROTECTED]">news:[EMAIL PROTECTED]...
> Leonardo Curros Berm�dez wrote:
>
> >Hello,
> >i have configured two instantces of NDS 6.11
> >and replicated both (master-master).
> >I have first used to initialice inicial server one ldif with 300.000
> >entries,
> >after i configure replica. At this moment searchs are fast (if i user
> >indexed attributes in filters).
> >
> >After this, I test it doing massive ldapdelete,ldapmodify and ldapadd.
> >Now when i do a query using a filter with indexed attributes, it takes so
> >long and
> >administrative timelimit is passed. It seems like index don�t regenerate.
> >Is this normal?
> >
> No. The indexes definitely regenerate. You should probably check your
> access log to see what the etime (elapsed time) is for your searches,
> and also check the notes to see if you are running into the all IDS
> threshold or doing unindexed searches.
>
> >
> >thanks
> >
> >
> >
> >
>
