Fedora directory users - Jul 2007 - Problem with Massive Replication Agreements

Hi all,

I have a distributed Directory Service on 25 servers with FDS 1.0.4. The
replication configuration it''s some complex:

- One HUB server with 26 Databases(26 enabled Replicas)-> 25
HUB-Replicas with 24 replication agreements and one Multimaster-Replica
with 26 replication agreements. Total replication agreements in this
server it''s 626 replication agreements.

- 25 Supplier (on our database) with Hub server and consummers (for
others databases) of Hub server. 

When i try to configure (automatically by script) this 626 replication
agreements (on HUB server) i show this
errorlog(/opt/fedora-ds/slapd-prueba/logs/errors) for each new
replication agreement after 250 replication agreements aprox.:

[22/Jul/2007:20:55:52 +0200] NSMMReplicationPlugin -
agmt="cn=prueba1-prueba2" (prueba1:636): Unable to create protocol
thread; NSPR error - -5974, Insufficient system resources.

Exist some FDS limitation about massive replication agreements? or some
FDS performance settings (maximun entries, caché,look-through limit,
size limit, file descriptors,connection management...)?

How can i solved this problem?

thanks,

The message tells you that the server failed to create a thread
because the OS wouldn''t let it. Typically this happens either because
there''s some OS limit that needs to be increased, or because the
thread stack stride is very large and you''re running a 32-bit server
(the fix for that is to reduce the stack size).

Fernando Muñoz wrote:
> Hi all,
>
> I have a distributed Directory Service on 25 servers with FDS 1.0.4. The
> replication configuration it''s some complex:
>
> - One HUB server with 26 Databases(26 enabled Replicas)-> 25
> HUB-Replicas with 24 replication agreements and one Multimaster-Replica
> with 26 replication agreements. Total replication agreements in this
> server it''s 626 replication agreements.
>
> - 25 Supplier (on our database) with Hub server and consummers (for
> others databases) of Hub server. 
>
> When i try to configure (automatically by script) this 626 replication
> agreements (on HUB server) i show this
> errorlog(/opt/fedora-ds/slapd-prueba/logs/errors) for each new
> replication agreement after 250 replication agreements aprox.:
>
> [22/Jul/2007:20:55:52 +0200] NSMMReplicationPlugin -
> agmt="cn=prueba1-prueba2" (prueba1:636): Unable to create
protocol
> thread; NSPR error - -5974, Insufficient system resources.
>
> Exist some FDS limitation about massive replication agreements? or some
> FDS performance settings (maximun entries, caché,look-through limit,
> size limit, file descriptors,connection management...)?
>
> How can i solved this problem?
>
> thanks,
>
>
>
>  
>
> --
> Fedora-directory-users mailing list
> Fedora-directory-users@redhat.com
> https://www.redhat.com/mailman/listinfo/fedora-directory-users
>

I''ve installed FDS over Centos 4.5.

How can i modify this OS limits or thread stack in Centos (RedHat)? and
then probe the replication configuration.

thanks,   

El dom, 22-07-2007 a las 11:09 -0600, David Boreham
escribió:
> The message tells you that the server failed to create a thread
> because the OS wouldn''t let it. Typically this happens either
because
> there''s some OS limit that needs to be increased, or because the
> thread stack stride is very large and you''re running a 32-bit
server
> (the fix for that is to reduce the stack size).
> 
> Fernando Muñoz wrote:
> > Hi all,
> >
> > I have a distributed Directory Service on 25 servers with FDS 1.0.4.
The
> > replication configuration it''s some complex:
> >
> > - One HUB server with 26 Databases(26 enabled Replicas)-> 25
> > HUB-Replicas with 24 replication agreements and one
Multimaster-Replica
> > with 26 replication agreements. Total replication agreements in this
> > server it''s 626 replication agreements.
> >
> > - 25 Supplier (on our database) with Hub server and consummers (for
> > others databases) of Hub server. 
> >
> > When i try to configure (automatically by script) this 626 replication
> > agreements (on HUB server) i show this
> > errorlog(/opt/fedora-ds/slapd-prueba/logs/errors) for each new
> > replication agreement after 250 replication agreements aprox.:
> >
> > [22/Jul/2007:20:55:52 +0200] NSMMReplicationPlugin -
> > agmt="cn=prueba1-prueba2" (prueba1:636): Unable to create
protocol
> > thread; NSPR error - -5974, Insufficient system resources.
> >
> > Exist some FDS limitation about massive replication agreements? or
some
> > FDS performance settings (maximun entries, caché,look-through limit,
> > size limit, file descriptors,connection management...)?
> >
> > How can i solved this problem?
> >
> > thanks,
> >
> >
> >
> >  
> >
> > --
> > Fedora-directory-users mailing list
> > Fedora-directory-users@redhat.com
> > https://www.redhat.com/mailman/listinfo/fedora-directory-users
> >   
> 
> --
> Fedora-directory-users mailing list
> Fedora-directory-users@redhat.com
> https://www.redhat.com/mailman/listinfo/fedora-directory-users

I solved the problem,

I reduced the stack size to 1024 (it was 10240) with ulimit -s 1024.

Thanks David,


El dom, 22-07-2007 a las 23:46 +0200, Fernando Muñoz
escribió:
> I''ve installed FDS over Centos 4.5.
> 
> How can i modify this OS limits or thread stack in Centos (RedHat)? and
> then probe the replication configuration.
> 
> thanks,   
> 
> El dom, 22-07-2007 a las 11:09 -0600, David Boreham escribió:
> > The message tells you that the server failed to create a thread
> > because the OS wouldn''t let it. Typically this happens either
because
> > there''s some OS limit that needs to be increased, or because
the
> > thread stack stride is very large and you''re running a 32-bit
server
> > (the fix for that is to reduce the stack size).
> > 
> > Fernando Muñoz wrote:
> > > Hi all,
> > >
> > > I have a distributed Directory Service on 25 servers with FDS
1.0.4. The
> > > replication configuration it''s some complex:
> > >
> > > - One HUB server with 26 Databases(26 enabled Replicas)-> 25
> > > HUB-Replicas with 24 replication agreements and one
Multimaster-Replica
> > > with 26 replication agreements. Total replication agreements in
this
> > > server it''s 626 replication agreements.
> > >
> > > - 25 Supplier (on our database) with Hub server and consummers
(for
> > > others databases) of Hub server. 
> > >
> > > When i try to configure (automatically by script) this 626
replication
> > > agreements (on HUB server) i show this
> > > errorlog(/opt/fedora-ds/slapd-prueba/logs/errors) for each new
> > > replication agreement after 250 replication agreements aprox.:
> > >
> > > [22/Jul/2007:20:55:52 +0200] NSMMReplicationPlugin -
> > > agmt="cn=prueba1-prueba2" (prueba1:636): Unable to
create protocol
> > > thread; NSPR error - -5974, Insufficient system resources.
> > >
> > > Exist some FDS limitation about massive replication agreements?
or some
> > > FDS performance settings (maximun entries, caché,look-through
limit,
> > > size limit, file descriptors,connection management...)?
> > >
> > > How can i solved this problem?
> > >
> > > thanks,
> > >
> > >
> > >
> > >  
> > >
> > > --
> > > Fedora-directory-users mailing list
> > > Fedora-directory-users@redhat.com
> > > https://www.redhat.com/mailman/listinfo/fedora-directory-users
> > >   
> > 
> > --
> > Fedora-directory-users mailing list
> > Fedora-directory-users@redhat.com
> > https://www.redhat.com/mailman/listinfo/fedora-directory-users
> 
> --
> Fedora-directory-users mailing list
> Fedora-directory-users@redhat.com
> https://www.redhat.com/mailman/listinfo/fedora-directory-users

when mattering the copied certificate of the servant in the data base of the
certificate using pk12util.exe.

I hava a problem:

C:\Archivos de programa\Red Hat Directory Password
Synchronization>pk12util.exe -d . -i adminserver.p12
Enter Password or Pin for "NSS Certificate DB":
Enter password for PKCS12 file:
pk12util.exe: PKCS12 decode not verified: The security password entered is
incorrect.


Thanks



Esteban Torres Rodríguez
ÁREA DE SOPORTE TÉCNICO - Administración de Servidores
Subdirección de Sistemas Informáticos
Empresa Pública Desarrollo Agrario y Pesquero, 
email: etorres@dap.es