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

Re: Questions about slapd & slurpd

At 04:39 PM 7/12/99 +0200, Frédéric Poels wrote:
>Good afternoon!
>Here are a few questions about slapd and slurpd for which I cannot find an
>answer in either tha FAQ or the "administrator's guide":
>1) Can slapd use a plain text logfile in a specified directory (I don't
>want to rely on syslogd)?

With a little hacking, yes.  Out of the box, no.

>2) Is it possible do have multiple referrals (one for a specific subtree
>and another one for another subtree)?


If you have a suffix "o=org" and three entries under this suffix,
	dn: o=org
	dn: l=left,o=org
	dn: l=right,o=org

You can then create two referral entries:
	dn: ref=ldap://left/,l=left,o=org
	dn: ref=ldap://right/,l=right,o=org

Then a subtree search with base "o=org" should return the three
original entries and a result that contains the two referrals.
If the base as "l=left,o=org", than only the left entry and referrral
would be returned on a subtree search.

>3) How can you perform a "total replication" (like X.500 allows)?

Using slurpd, you must replicate the data assocatiated with each
"database" separately.

>Copying DB files (as described in the administrator's guide) just
>does not work between heterogeneous systems (Linux and AIX to name them).

Even between homogeneous systems, you shouldn't copy the database files
between the master and it's slaves.

>What can you do
>if the "supplier" uses DB2 and the "consumer" uses GDBM? What can you do if
>you do not have file transfer facilities between the two systems?

Use slurpd.

>[ I'm using Sleepycat Berkeley DB2 and db_dump/db_load don't seem to help
>much ]
>Any hint is welcome!
>Frederic Poels.