[Fedora-directory-users] Glue Entry Thread

[James]

Hi All,

I have a set of directory servers with multi-master replicaiton.  On one of 
the two master servers, I see this log:

[25/Mar/2008:14:26:42 -0400] NSMMReplicationPlugin - conn=5 op=6 
csn=47cec1700000000c0000:
 Can't created glue entry 
uid=soleotester,ou=people,dc=soleocommunications,dc=com uniqueid
=96a7eb81-1dd111b2-8016d669-d3980000, error 68
[25/Mar/2008:14:26:42 -0400] NSMMReplicationPlugin - conn=5 op=6 
csn=47cec1700000000c0000:
 Can't created glue entry 
uid=soleotester,ou=people,dc=soleocommunications,dc=com uniqueid
=96a7eb81-1dd111b2-8016d669-d3980000, error 68

The logs is repeated once per second (there are two in this copy/paste).  I 
have a high-level understanding of what a glue entry is, and why one would be 
created, but why can't this server create one in this instance?  And, is 
there anything I can do to fix this repeated log?

Thanks,
~James


-- 
James Bushey
Software Engineer
Soleo Communications
[phone] x0050

[Nathan Kinder]

James wrote:
>  Hi All,
>
>  I have a set of directory servers with multi-master replicaiton.  On one of 
>  the two master servers, I see this log:
>
>  [25/Mar/2008:14:26:42 -0400] NSMMReplicationPlugin - conn=5 op=6 
>  csn=47cec1700000000c0000:
>   Can't created glue entry 
>  uid=soleotester,ou=people,dc=soleocommunications,dc=com uniqueid
>  =96a7eb81-1dd111b2-8016d669-d3980000, error 68
>  [25/Mar/2008:14:26:42 -0400] NSMMReplicationPlugin - conn=5 op=6 
>  csn=47cec1700000000c0000:
>   Can't created glue entry 
>  uid=soleotester,ou=people,dc=soleocommunications,dc=com uniqueid
>  =96a7eb81-1dd111b2-8016d669-d3980000, error 68
>
>  The logs is repeated once per second (there are two in this copy/paste).  I 
>  have a high-level understanding of what a glue entry is, and why one would be 
>  created, but why can't this server create one in this instance?  And, is 
>  there anything I can do to fix this repeated log?
>     
It can't create it because it already exists (error 68).  Please file a 
bug on this issue (https://bugzilla.redhat.com/enter_bug.cgi).

You can try to delete the existing glue entry to allow the replication 
plug-in to re-create it and proceed.

-NGK
>  Thanks,
>  ~James
>
>
>

[James]

Thanks for the suggestion.  I have tried searching for the glue entry in the 
database, and I cant find it:

[soleo@mstrldap01 ~]$ ldapsearch -MMxw xxxxx -D "cn=Directory 
Manager" -b "ou=soleotester,ou=people,dc=soleocommunications,dc=com" -s 
one -h 10.1.5.211
# extended LDIF
#
# LDAPv3
# base <ou=soleotester,ou=people,dc=soleocommunications,dc=com> with scope one
# filter: (objectclass=*)
# requesting: ALL
# with manageDSAit critical control
#

# search result
search: 2
result: 32 No such object
matchedDN: ou=people,dc=soleocommunications,dc=com

# numResponses: 1

When I first noticed these logs, I did find the original entry present on this 
server (and on the other master) so I deleted this entry from both servers 
(and restarted ns-slapd), but that didnt get rid of the log.

Also, Ive noticed that after a while of having this error printed out, the 
server stops allowing me to bind in.

Am I doing something wrong in my search?  Or, is there something else I can 
try?

Thanks

~James

On Tuesday 25 March 2008 14:46:56 Nathan Kinder wrote:
>  James wrote:
>  > Hi All,
>  >
>  > I have a set of directory servers with multi-master replicaiton.  On one
>  > of the two master servers, I see this log:
>  >
>  > [25/Mar/2008:14:26:42 -0400] NSMMReplicationPlugin - conn=5 op=6
>  > csn=47cec1700000000c0000:
>  >  Can't created glue entry
>  > uid=soleotester,ou=people,dc=soleocommunications,dc=com uniqueid
>  > =96a7eb81-1dd111b2-8016d669-d3980000, error 68
>  > [25/Mar/2008:14:26:42 -0400] NSMMReplicationPlugin - conn=5 op=6
>  > csn=47cec1700000000c0000:
>  >  Can't created glue entry
>  > uid=soleotester,ou=people,dc=soleocommunications,dc=com uniqueid
>  > =96a7eb81-1dd111b2-8016d669-d3980000, error 68
>  >
>  > The logs is repeated once per second (there are two in this copy/paste). 
>  > I have a high-level understanding of what a glue entry is, and why one
>  > would be created, but why can't this server create one in this instance? 
>  > And, is there anything I can do to fix this repeated log?
>
>  It can't create it because it already exists (error 68).  Please file a
>  bug on this issue (https://bugzilla.redhat.com/enter_bug.cgi).
>
>  You can try to delete the existing glue entry to allow the replication
>  plug-in to re-create it and proceed.
>
>  -NGK
>
>  > Thanks,
>  > ~James 


-- 
James Bushey
Software Engineer
Soleo Communications
[phone] x0050

[Nathan Kinder]

James wrote:
>  Thanks for the suggestion.  I have tried searching for the glue entry
> in the 
>  database, and I cant find it:
>
>  [soleo@mstrldap01 ~]$ ldapsearch -MMxw xxxxx -D "cn=Directory 
>  Manager" -b "ou=soleotester,ou=people,dc=soleocommunications,dc=com" -s 
>  one -h 10.1.5.211
>  # extended LDIF
>  #
>  # LDAPv3
>  # base <ou=soleotester,ou=people,dc=soleocommunications,dc=com> with scope one
>  # filter: (objectclass=*)
>  # requesting: ALL
>  # with manageDSAit critical control
>  #
>
>  # search result
>  search: 2
>  result: 32 No such object
>  matchedDN: ou=people,dc=soleocommunications,dc=com
>
>  # numResponses: 1
>
>  When I first noticed these logs, I did find the original entry present on this 
>  server (and on the other master) so I deleted this entry from both servers 
>  (and restarted ns-slapd), but that didnt get rid of the log.
>
>  Also, Ive noticed that after a while of having this error printed out, the 
>  server stops allowing me to bind in.
>
>  Am I doing something wrong in my search?  Or, is there something else I can 
>  try?
>     
Your search is searching for 
"ou=soleotester,ou=people,dc=soleocommunications,dc=com", but the glue 
entry the server is trying to create is 
"uid=soleotester,ou=people,dc=soleocommunications,dc=com".  Try doing 
this search instead:

  ldapsearch -b "ou=people,dc=soleocommunications,dc=com" -s one 
"uid=soleotester"

-NGK


>  Thanks
>
>  ~James
>
>  On Tuesday 25 March 2008 14:46:56 Nathan Kinder wrote:
>    
> > James wrote:
> >     
> >> Hi All,
> >>
> >> I have a set of directory servers with multi-master replicaiton.  On one
> >> of the two master servers, I see this log:
> >>
> >> [25/Mar/2008:14:26:42 -0400] NSMMReplicationPlugin - conn=5 op=6
> >> csn=47cec1700000000c0000:
> >>  Can't created glue entry
> >> uid=soleotester,ou=people,dc=soleocommunications,dc=com uniqueid
> >> =96a7eb81-1dd111b2-8016d669-d3980000, error 68
> >> [25/Mar/2008:14:26:42 -0400] NSMMReplicationPlugin - conn=5 op=6
> >> csn=47cec1700000000c0000:
> >>  Can't created glue entry
> >> uid=soleotester,ou=people,dc=soleocommunications,dc=com uniqueid
> >> =96a7eb81-1dd111b2-8016d669-d3980000, error 68
> >>
> >> The logs is repeated once per second (there are two in this copy/paste). 
> >> I have a high-level understanding of what a glue entry is, and why one
> >> would be created, but why can't this server create one in this instance? 
> >> And, is there anything I can do to fix this repeated log?
> >>       
> > It can't create it because it already exists (error 68).  Please file a
> > bug on this issue (https://bugzilla.redhat.com/enter_bug.cgi).
> >
> > You can try to delete the existing glue entry to allow the replication
> > plug-in to re-create it and proceed.
> >
> > -NGK
> >
> >     
> >> Thanks,
> >> ~James
> >>       
>
>
>
>

[James]

Whoop!  I ran that search in a hurry yesterday just to have something to paste 
into the email.  I had run my search and the one you suggested earlier 
(correctly) but couldnt see the entry.

I did, however, manage to delete the entry which I could not see.  The trick 
was, I had to shut down both masters, then start only the master that had 
been printing out the errors.  At this point, the error-printing master was 
not printing the error (no replication from the other server pushing this 
change across).  Then, I was able to do an ldapmodify and delete the 
invisible glue entry.  I then started up the other master, and replication 
resumed as normal, without any error messages.

That seemed to work for me.  Thanks for all of your help.

~James

[soleo@mstrldap01 ~]$ ldapsearch -MMxw xxxx -D "cn=Directory 
Manager" -b "ou=people,dc=soleocommunications,dc=com" -s one -h 
10.1.5.211 '(uid=soleotester)'
# extended LDIF
#
# LDAPv3
# base <ou=people,dc=soleocommunications,dc=com> with scope one
# filter: (uid=soleotester)
# requesting: ALL
# with manageDSAit critical control
#

# search result
search: 2
result: 0 Success

# numResponses: 1


On Tuesday 25 March 2008 18:26:26 Nathan Kinder wrote:
>  James wrote:
>  > Thanks for the suggestion.  I have tried searching for the glue entry in
>  > the database, and I cant find it:
>  >
>  > [soleo@mstrldap01 ~]$ ldapsearch -MMxw xxxxx -D "cn=Directory
>  > Manager" -b "ou=soleotester,ou=people,dc=soleocommunications,dc=com"
> -s
>  > one -h 10.1.5.211
>  > # extended LDIF
>  > #
>  > # LDAPv3
>  > # base <ou=soleotester,ou=people,dc=soleocommunications,dc=com> with
>  > scope one # filter: (objectclass=*)
>  > # requesting: ALL
>  > # with manageDSAit critical control
>  > #
>  >
>  > # search result
>  > search: 2
>  > result: 32 No such object
>  > matchedDN: ou=people,dc=soleocommunications,dc=com
>  >
>  > # numResponses: 1
>  >
>  > When I first noticed these logs, I did find the original entry present on
>  > this server (and on the other master) so I deleted this entry from both
>  > servers (and restarted ns-slapd), but that didnt get rid of the log.
>  >
>  > Also, Ive noticed that after a while of having this error printed out,
>  > the server stops allowing me to bind in.
>  >
>  > Am I doing something wrong in my search?  Or, is there something else I
>  > can try?
>
>  Your search is searching for
>  "ou=soleotester,ou=people,dc=soleocommunications,dc=com", but the glue
>  entry the server is trying to create is
>  "uid=soleotester,ou=people,dc=soleocommunications,dc=com".  Try doing
>  this search instead:
>
>    ldapsearch -b "ou=people,dc=soleocommunications,dc=com" -s one
>  "uid=soleotester"
>
>  -NGK
>
>  > Thanks
>  >
>  > ~James
>  >
>  > On Tuesday 25 March 2008 14:46:56 Nathan Kinder wrote:
>  >> James wrote:
>  >>> Hi All,
>  >>>
>  >>> I have a set of directory servers with multi-master replicaiton.  On
>  >>> one of the two master servers, I see this log:
>  >>>
>  >>> [25/Mar/2008:14:26:42 -0400] NSMMReplicationPlugin - conn=5 op=6
>  >>> csn=47cec1700000000c0000:
>  >>>  Can't created glue entry
>  >>> uid=soleotester,ou=people,dc=soleocommunications,dc=com uniqueid
>  >>> =96a7eb81-1dd111b2-8016d669-d3980000, error 68
>  >>> [25/Mar/2008:14:26:42 -0400] NSMMReplicationPlugin - conn=5 op=6
>  >>> csn=47cec1700000000c0000:
>  >>>  Can't created glue entry
>  >>> uid=soleotester,ou=people,dc=soleocommunications,dc=com uniqueid
>  >>> =96a7eb81-1dd111b2-8016d669-d3980000, error 68
>  >>>
>  >>> The logs is repeated once per second (there are two in this
>  >>> copy/paste). I have a high-level understanding of what a glue entry is,
>  >>> and why one would be created, but why can't this server create one in
>  >>> this instance? And, is there anything I can do to fix this repeated
>  >>> log?
>  >>
>  >> It can't create it because it already exists (error 68).  Please file a
>  >> bug on this issue (https://bugzilla.redhat.com/enter_bug.cgi).
>  >>
>  >> You can try to delete the existing glue entry to allow the replication
>  >> plug-in to re-create it and proceed.
>  >>
>  >> -NGK
>  >>
>  >>> Thanks,
>  >>> ~James 


-- 
James Bushey
Software Engineer
Soleo Communications
[phone] x0050

[Carlos Barrales Ruiz]

El 25/03/2008, a las 19:28, James escribió:

>  Hi All,
>
>  I have a set of directory servers with multi-master replicaiton.  On  
>  one of
>  the two master servers, I see this log:
>
>  [25/Mar/2008:14:26:42 -0400] NSMMReplicationPlugin - conn=5 op=6
>  csn=47cec1700000000c0000:
>  Can't created glue entry
>  uid=soleotester,ou=people,dc=soleocommunications,dc=com uniqueid
>  =96a7eb81-1dd111b2-8016d669-d3980000, error 68
>  [25/Mar/2008:14:26:42 -0400] NSMMReplicationPlugin - conn=5 op=6
>  csn=47cec1700000000c0000:
>  Can't created glue entry
>  uid=soleotester,ou=people,dc=soleocommunications,dc=com uniqueid
>  =96a7eb81-1dd111b2-8016d669-d3980000, error 68
>
>  The logs is repeated once per second (there are two in this copy/ 
>  paste).  I
>  have a high-level understanding of what a glue entry is, and why one  
>  would be
>  created, but why can't this server create one in this instance?   
>  And, is
>  there anything I can do to fix this repeated log? 
Hi James,

We had the same problem. It seemed to me that mmr plugin could not  
create the entry because it was created and specified to be a root-of- 
replication suffix after it (or another suffix parented to it) was  
still being replicated.

So, in our case, as we had the directory not populated... recreating  
the tree before starting replication was enougth for us to solve it.

Maybe you should get the sucessfull behavior removing manually the  
changelog, reseting the replication states that affect the non- 
creatable-entry and reinitializing them, but no warranty. I suppose  
also that it should be fixed in incremental way by removing the  
affected entry in the server causing the error (consumer) and force  
the replication supplier to populate it.

I'm not sure either on if problem comes from the local integrity  
caused by the unique entry id of the entry.

Sorry for my bad english.

Regards
--
Carlos