Hi Maikel,

the conninfo in all paths in create_new.sh is the same. It should point to
different servers like in replication.sh.

regards

Hendrik Woltersdorf
XCOM AG
Tel.: 0375/27008-162
Tel. mobil: 0178/5710078
Email: Hendrik.Woltersdorf at xcom.de


                                                                           
             Maikel Verheijen                                              
             <maikel.verheijen                                             
             @active24.nl>                                              An 
             Gesendet von:              slony1-general at gborg.postgresql.or 
             slony1-general-bo          g                                  
             unces at gborg.postg                                       Kopie 
             resql.org                                                     
                                                                     Thema 
                                        [Slony1-general] replication       
             29.10.2004 07:58           problem (newby question?) [auf     
                                        Viren geprueft]                    
                                                                           
                                                                           
                                                                           
                                                                           
                                                                           
                                                                           




Hi list,

While trying to re-create a 3-node database-cluster, I started creating
the 3 node cluster using the example of the pg_bench, and extended it
with 1 node. The replication worked perfectly, I inserted 650.000 lines
in one table, and 65.000 in another, and the replicated servers didn't
fall back much while I was inserting.

Querying the 3 databases resulted in the same data, so I was confident
it all worked ok.

My next test was to remove a node, and recreate it. The deletion seemed
to go OK, and the recreate seemed to go OK, but when my slon is
started, I get these errors:

Node1: (the master)
ERROR  remoteListenThread_2: db_getLocalNodeId() returned 1 - wrong
database?

Node2: (the deleted and re-created slave)
ERROR  remoteListenThread_3: db_getLocalNodeId() returned 1 - wrong
database?

Node3: (the slave that kept working after the delete)
ERROR  remoteListenThread_2: db_getLocalNodeId() returned 1 - wrong
database?

I hope anyone can point me in the right direction, I would really love
to solve this issue :)

My create, start, delete and re-create scripts are attached. The tables
are a derivate of the powerdns

Thank you in advance,

Maikel Verheijen

Ps: My passwords and IP's are changed to different values
Pps: the run.sh script is changed per-node. (the THISHOST and THISDB
variables are changed)


[Anhang "create_new.sh" gel?scht von Hendrik Woltersdorf/DE/XCOM] [Anhang
"del_replication.sh" gel?scht von Hendrik Woltersdorf/DE/XCOM] [Anhang
"replication_start.sh" gel?scht von Hendrik Woltersdorf/DE/XCOM] [Anhang
"replication.sh" gel?scht von Hendrik Woltersdorf/DE/XCOM] [Anhang "run.sh"
gel?scht von Hendrik Woltersdorf/DE/XCOM] [Anhang "PGP.sig" gel?scht von
Hendrik Woltersdorf/DE/XCOM]
_______________________________________________
Slony1-general mailing list
Slony1-general at gborg.postgresql.org
http://gborg.postgresql.org/mailman/listinfo/slony1-general