Re: slapadd -q (quick mode)

[Howard Chu <hyc@symas.com>]

Hallvard B Furuseth wrote:
>  Howard Chu writes:
> >> Though actually I would like an option for more LDIF error
> >> checking too: Adding whatever check ldapadd does which is
> >> convenient to add to slapadd, including a "verify" function in
> >> overlays like "unique", to be called for each entry after the
> >> database is built and indexed.
> > Then I think you should just use ldapadd in the first place.
>
>  Well... To get the same speed that would mean to run a separate slapd
>  without logs and without consistency checks.  Can DB_CONFIG be used
>  to configure slapd database operations to be more or less as fast as
>  slapadd -q?

Pretty close. If you stick with BDB 4.2 you can disable txn logs and 
synchronous writes, but slapd always runs with txn's enabled.

>  Does slapadd still offer a speedup by adding entries first and
>  indexing the database later?  ldapadd can't do that.

slapadd doesn't do that, it adds indexes for an entry right when the 
entry is added, same as ldapadd. (Unless you're using the back-bdb 
linearindex option, but that in fact is slower when your BDB cache is 
large enough.)

>  Will ldapadd fail as reliably as slapadd if a database error happens?

The same code paths are followed for the most part, so I don't see why not.

--
 -- Howard Chu
 Chief Architect, Symas Corp.  http://www.symas.com
 Director, Highland Sun        http://highlandsun.com/hyc
 OpenLDAP Core Team            http://www.openldap.org/project/