Re: replication having multiple backends

["Pierangelo Masarati" <ando@sys-net.it>]

> hi there!
>
> when looking at the slurpd source code, I found out, that the suffix
> statement in the replica context is ignored:

as intended.

>
> excerpt from config.c:
>
>          } else if ( !strncasecmp( cargv[ i ],
>                          SUFFIXSTR, sizeof( SUFFIXSTR ) - 1 ) ) {
>              /* ignore it */ ;
>
> so hopefully the programmers will activate this statement soon, and in
> that way I feel it is intended to, namely to make it possible to do the
> replication for multiple backends hooked on different branches, without
> getting referral errors back from the slave slapd.

It is ignored because it is slapd's responsibility to schedule
replicas for the approriate slurpd threads; those, on turn,
simply collect their chenages and propagate them to the server.
If you look at the corresponding slapd's code, it is not ignored.

> *waving hands to the programmers corner* :)
>
> and please be so kind and document not implemented stuff correctly in
> the man pages.

The documentation correctly describes what the software does
and it needs no changes at all.  Please read the code and the
documentation carefully, and you'll find out that they match.

p.

-- 
Pierangelo Masarati
mailto:pierangelo.masarati@sys-net.it