[Fedora-directory-users] Replication agreement trouble

Wed Apr 22 06:38:51 UTC 2009

The day before the date in the error (when the errors started), we we
had to delete two suffix databases from the console (they were damaged),
create them again, and reinitialize those databases from other supplier.
The database of the agreement throwing errors is the userRoot
(dc=example,dc=com). The databases recreated were the suffixes
o=cabu,dc=sacyl,dc=es and o=husa,dc=sacyl,dc=es.

This is the error log from server1 (this did not crash, this server
initialized the server2, that crashed):

===========================

[20/Apr/2009:14:18:28 +0200] NSMMReplicationPlugin - Beginning total
update of replica "agmt="cn=CABU_ppal-GRS_back" (grsgscvalp0102:636)".
[20/Apr/2009:14:18:39 +0200] NSMMReplicationPlugin - Finished total
update of replica "agmt="cn=CABU_ppal-GRS_back" (grsgscvalp0102:636)".
Sent 4108 entries.
[20/Apr/2009:14:25:33 +0200] NSMMReplicationPlugin - Beginning total
update of replica "agmt="cn=HUSA_ppal-GRS_back" (grsgscvalp0102:636)".
[20/Apr/2009:14:25:43 +0200] NSMMReplicationPlugin - Finished total
update of replica "agmt="cn=HUSA_ppal-GRS_back" (grsgscvalp0102:636)".
Sent 2650 entries.
[21/Apr/2009:10:50:47 +0200] - slapd shutting down - signaling operation
threads

===========================

And this is the log from server2, where the databases crashed. The log
shows the deletion of the agreements, the deletion of the databases, the
creation of the databases and the initialization of them from server1.
The messages from day 21 are when we tried to force to send the updates:

===========================

[20/Apr/2009:14:13:20 +0200] NSMMReplicationPlugin - agmt_delete: begin
[20/Apr/2009:14:13:21 +0200] NSMMReplicationPlugin -
multimaster_be_state_change: replica dc=sacyl,dc=es is about to be
deleted; disabling replication
[20/Apr/2009:14:14:16 +0200] - ldbm: Bringing o_cabu_dc_sacyl_dc_es
offline...
[20/Apr/2009:14:14:16 +0200] - ldbm: removing 'o_cabu_dc_sacyl_dc_es'.
[20/Apr/2009:14:14:16 +0200] - Destructor for instance
o_cabu_dc_sacyl_dc_es called
[20/Apr/2009:14:14:44 +0200] - No symmetric key found for cipher AES in
backend o_cabu_dc_sacyl_dc_es, attempting to create one...
[20/Apr/2009:14:14:44 +0200] - Key for cipher AES successfully generated
and stored
[20/Apr/2009:14:14:44 +0200] - No symmetric key found for cipher 3DES in
backend o_cabu_dc_sacyl_dc_es, attempting to create one...
[20/Apr/2009:14:14:45 +0200] - Key for cipher 3DES successfully
generated and stored
[20/Apr/2009:14:17:08 +0200] NSMMReplicationPlugin -
agmt="cn=CABU_back-GRS_ppal" (grsgscvalp0101:636): Replica has a
different generation ID than the local 
data.
[20/Apr/2009:14:18:11 +0200] NSMMReplicationPlugin -
multimaster_be_state_change: replica o=cabu,dc=sacyl,dc=es is going
offline; disabling replication
[20/Apr/2009:14:18:13 +0200] - WARNING: Import is running with
nsslapd-db-private-import-mem on; No other process is allowed to access
the database
[20/Apr/2009:14:18:35 +0200] - import o_cabu_dc_sacyl_dc_es: Workers
finished; cleaning up...
[20/Apr/2009:14:18:36 +0200] - import o_cabu_dc_sacyl_dc_es: Workers
cleaned up.
[20/Apr/2009:14:18:36 +0200] - import o_cabu_dc_sacyl_dc_es: Indexing
complete.  Post-processing...
[20/Apr/2009:14:18:36 +0200] - import o_cabu_dc_sacyl_dc_es: Flushing
caches...
[20/Apr/2009:14:18:36 +0200] - import o_cabu_dc_sacyl_dc_es: Closing
files...
[20/Apr/2009:14:18:38 +0200] - import o_cabu_dc_sacyl_dc_es: Import
complete.  Processed 4108 entries in 12 seconds. (342.33 entries/sec)
[20/Apr/2009:14:18:39 +0200] NSMMReplicationPlugin -
multimaster_be_state_change: replica o=cabu,dc=sacyl,dc=es is coming
online; enabling replication
[20/Apr/2009:14:20:09 +0200] NSMMReplicationPlugin -
replica_config_delete: Warning: The changelog for replica
o=husa,dc=sacyl,dc=es is no longer valid since
 the replica config is being deleted.  Removing the changelog.
[20/Apr/2009:14:20:10 +0200] NSMMReplicationPlugin - agmt_delete: begin
[20/Apr/2009:14:20:12 +0200] NSMMReplicationPlugin -
multimaster_be_state_change: replica dc=sacyl,dc=es is about to be
deleted; disabling replication
[20/Apr/2009:14:20:42 +0200] - ldbm: Bringing o_husa_dc_sacyl_dc_es
offline...
[20/Apr/2009:14:20:42 +0200] - ldbm: removing 'o_husa_dc_sacyl_dc_es'.
[20/Apr/2009:14:20:42 +0200] - Destructor for instance
o_husa_dc_sacyl_dc_es called
[20/Apr/2009:14:21:10 +0200] - No symmetric key found for cipher AES in
backend o_husa_dc_sacyl_dc_es, attempting to create one...
[20/Apr/2009:14:21:10 +0200] - Key for cipher AES successfully generated
and stored
[20/Apr/2009:14:21:10 +0200] - No symmetric key found for cipher 3DES in
backend o_husa_dc_sacyl_dc_es, attempting to create one...
[20/Apr/2009:14:21:10 +0200] - Key for cipher 3DES successfully
generated and stored
[20/Apr/2009:14:24:23 +0200] NSMMReplicationPlugin -
agmt="cn=HUSA_back-GRS_ppal" (grsgscvalp0101:636): Replica has a
different generation ID than the local 
data.
[20/Apr/2009:14:25:18 +0200] NSMMReplicationPlugin -
multimaster_be_state_change: replica o=husa,dc=sacyl,dc=es is going
offline; disabling replication
[20/Apr/2009:14:25:20 +0200] - WARNING: Import is running with
nsslapd-db-private-import-mem on; No other process is allowed to access
the database
[20/Apr/2009:14:25:39 +0200] - import o_husa_dc_sacyl_dc_es: Workers
finished; cleaning up...
[20/Apr/2009:14:25:40 +0200] - import o_husa_dc_sacyl_dc_es: Workers
cleaned up.
[20/Apr/2009:14:25:40 +0200] - import o_husa_dc_sacyl_dc_es: Indexing
complete.  Post-processing...
[20/Apr/2009:14:25:40 +0200] - import o_husa_dc_sacyl_dc_es: Flushing
caches...
[20/Apr/2009:14:25:40 +0200] - import o_husa_dc_sacyl_dc_es: Closing
files...
[20/Apr/2009:14:25:42 +0200] - import o_husa_dc_sacyl_dc_es: Import
complete.  Processed 2650 entries in 8 seconds. (331.25 entries/sec)
[20/Apr/2009:14:25:42 +0200] NSMMReplicationPlugin -
multimaster_be_state_change: replica o=husa,dc=sacyl,dc=es is coming
online; enabling replication
[21/Apr/2009:10:50:07 +0200] NSMMReplicationPlugin - Replication
agreement for agmt="cn=GRS_back-GRS_ppal" (grsgscvalp0101:636) could not
be updated. For rep
lication to take place, please enable the suffix and restart the server
[21/Apr/2009:10:50:07 +0200] NSMMReplicationPlugin - Replication
agreement for agmt="cn=GRS_back-GRS_ppal" (grsgscvalp0101:636) could not
be updated. For rep
lication to take place, please enable the suffix and restart the server

===========================

El mar, 21-04-2009 a las 09:21 -0600, Rich Megginson escribió:

> Juan Asensio Sánchez wrote:
> > Hi
> >
> > Since yesterday I am having troubles with replication between two 
> > servers. The replica is in multimaster mode in both servers, and 
> > everything is configured OK (database, suffixes, changelog, replica, 
> > agreements; until yesterday everything worked OK).
> >
> > [21/Apr/2009:11:04:57 +0200] NSMMReplicationPlugin - Replication 
> > agreement for agmt="cn=GRS_back-GRS_ppal" (grsgscvalp0101:636) could 
> > not be updated. For replication to take place, please enable the 
> > suffix and restart the server
> What changed?  Everything was working, then suddenly it's not?  
> Something must have changed, perhaps even something that did not seem 
> related to this problem.  Do you know when things started failing?  Did 
> you examine the access and error logs on the supplier and consumer from 
> around the time of the failure?
> >
> > The only thing to mention are replication problems with other 
> > databases and replicas, but not for the replica of the agreement in 
> > the message. They were fixed re-initializing the consumers of those 
> > replicas. Any idea?
> >
> > Regards and thanks in advance.
> > ------------------------------------------------------------------------
> >
> > --
> > Fedora-directory-users mailing list
> > Fedora-directory-users at redhat.com
> > https://www.redhat.com/mailman/listinfo/fedora-directory-users
> >   
> 
> 
> --
> Fedora-directory-users mailing list
> Fedora-directory-users at redhat.com
> 
> https://www.redhat.com/mailman/listinfo/fedora-directory-users
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.fedoraproject.org/pipermail/389-users/attachments/20090422/7688d01e/attachment.html>