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

Re: only one slurp per machine ?

On 10/19/06, Pierangelo Masarati <ando@sys-net.it> wrote:
> Hello,
> I'm testing a configuration with several master und replicas
> at one machine (2.3.25), but when there seems to be no way
> to start a second slurpd ?

You don't need a second slurpd, while you might want to use separate
replogs for each database.  Have a look at
<http://www.openldap.org/faq/data/cache/1171.html> for details, known
issues and suggested workarounds.

> Any Ideas or should I take syncrepl for this purpose ?

You should anyway, since slurpd is approaching end of life, it's been
essentially unsupported in 2.3, and won't be distributed any longer at
some point.

Despite all of this, here's how to do it:

In your different "database" definitions, simply use separate repolog
files.  Then start slurpds with -f slurpd.db1.conf -r replog.db1 -t
/path/to/db1slurpdlogdir/.  (the -r isn't needed if replog is defined
in slurpd.conf)

in backend.conf:
database bdb
suffix foo
include replication1.conf

database bdb
suffix bar
include replication2.conf

in replication.db1.conf:
replogfile replog.db1
host stuff credentials etc

in slurpd.db1.conf:
include replication1.conf
(note: do not include your backends here, or you'll get slurpd
replicating to all of the hosts mentioned)

If you're not spliiting things up by database, then this won't work.
(you probably don't gain anything -with- separate slurpd's anyway, but
please do some testing and let us know)