Hi Well, I dit solve this issue adding the URLs of the referrals to the list of referrals in the replica, not only in the suffix configuration. After this, those log messages did not appear anymore, and the referrals are returned correctly when trying to modify an entry that should return a referral. Hope this can help anyone. Regards. El día 26 de octubre de 2009 10:33, Juan Asensio Sánchez <okelet@xxxxxxxxx> escribió: > Hi again > > What's the difference between the referrals specified in the suffixes > and the referrals specified in a replica? We have had ever the > referrals specified in the suffix. Do you mean to set them too in the > replica? > > Regards. > > (Testing 389DS 1.2.3, having problems with schema syntaxes...) > > > 2009/10/19 Rich Megginson <rmeggins@xxxxxxxxxx>: >> Juan Asensio Sánchez wrote: >>> >>> Hi >>> >>> We are having problems with replication. We have four master servers >>> replicating one database (database 1), and two other servers in other >>> building that are masters of other database (database 2). Replication >>> between the databases of the others are in hub mode (server1 of >>> building1 with server1 of building1, and server2 of building1 with >>> server2 of building2; server3 and server4 of building1 only have >>> agreements with server1 and server2 of building1). Yesterday we had to >>> remove the replica of server4 of database 1. Next, we did ree-enable >>> the replica and the replication agreements, but the replica was >>> enabled with a different id than before. Now, we have this error on >>> all servers, although replication is working fine: >>> >>> [16/Oct/2009:12:44:39 +0200] NSMMReplicationPlugin - >>> repl_set_mtn_referrals: could not set referrals for replica >>> dc=domain,dc=local: 1 >>> >> >> I don't think this is a critical error. You can manually set the referrals >> to use - in the console for the Replica config, at the bottom of that window >> (under the Supplier DN area) you can specify the referrals to use. However, >> you will have to manage these manually - they are not set automatically like >> the regular replication referrals are. >> >>> (dc=domain,dc=local is the prefix that owns the database 1). I don't >>> know if this error is critical, but i don't like to see errors in the >>> log (call me fool if you want). I have read some posts: >>> >>> - http://blogs.sun.com/marcos/entry/on_cleanruv >>> - http://blogs.sun.com/piotr/entry/how_to_clean_ruv_s >>> >>> All about Sun, but i am not sure if this will work, or if it is >>> dangerous because it says that the solution is unsupported and >>> irreversible. How can I get rid of this message? is it critical? >>> >> >> You could try using that. I would suggest backing up all of your data and >> config first. >>> >>> Regards and thanks in advance. >>> >>> -- >>> 389 users mailing list >>> 389-users@xxxxxxxxxx >>> https://www.redhat.com/mailman/listinfo/fedora-directory-users >>> >> >> >> >> -- >> 389 users mailing list >> 389-users@xxxxxxxxxx >> https://www.redhat.com/mailman/listinfo/fedora-directory-users >> >> > -- 389 users mailing list 389-users@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-directory-users
