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

Re: slurpd



>
>Then perhaps your master server is using the wrong bind dn in it's replica statement?
>

No such luck.  Wouldn't that prevent the replication from occuring completely?  What I'm 
seeing seems like a bug, or at least undesirable behavior.

With a little more testing I've determined this is cause because I have 2 databases 
being replicated to the same machine like so:

# root suffix not replicated (only 1 entry)
dc=microcenter, dc=com

# information about a type of diskless terminal
# every one of the locations below have this info replicated to the location's server
ou=platforms, dc=microcenter, dc=com	#

# individual stores
# each one the following locations gets replicated to a specific server
location=co, dc=microcenter, dc=com	# -- copos.microcenter.com
location=ho, dc=microcenter, dc=com	# -- dbox.microcenter.com
location=tn, dc=microcenter, dc=com	# -- tnpos.microcenter.com
.
.					# 16 more locations
.

In the "ou=platforms,..." and the "location=ho,..." there is a line like:

replica	hostname=dbox.microcenter.com:389
	binddn="this varies depending which suffix" bindmethod=simple
	credentials=password
	
If I comment out one of the replica directives or comment out one of the databases on 
the slave and modify the remaining database it is replicated with out error.  It doesn't 
matter which suffix I get rid of and it doesn't matter whether it is on the master or 
slave.

Anthony

ps Thanks for reading so far :)