Le 26/02/2011 01:12, Hugo Pérez Casanova a écrit :
> Hi Everybody.
> 
> I'm not sure if this is the right list, correct me if I'm wrong.

Yes it is. There is no better place than slony mailing-lists to talk about
slony :-)

> 
> This is the first time I've worked with slony.
> 
> I made a test environment with two virtual servers working in CentOS 5.5, 
> Postgresql 8.4 and Slony 2.0.6. Test environment works fine.
> 
> When moving to production I have cloned a mid size database, when replication 
> starts I see at the slon log that it founds all tables to replicate. The 
> problem appears when writing to a tmp archive. Following is a snippet of the 
> log:
> 
> 2011-02-25 18:07:32 CSTCONFIG remoteWorkerThread_1: prepare to copy 
> table "public"."model_devices"
> 2011-02-25 18:07:32 CSTCONFIG remoteWorkerThread_1: prepare to copy 
> table "public"."array_name"
> 2011-02-25 18:07:32 CSTCONFIG remoteWorkerThread_1: all tables for set 1 found 
> on subscriber
> 2011-02-25 18:07:32 CSTCONFIG remoteWorkerThread_1: copy 
> table "public"."provider"
> 2011-02-25 18:07:33 CSTCONFIG remoteWorkerThread_1: Begin COPY of 
> table "public"."provider"
> NOTICE:  truncate of "public"."provider" failed - doing delete
> 2011-02-25 18:07:33 CSTERROR  remoteWorkerThread_1: Cannot write to archive 
> file /tmp/slony1_log_2_00000000000000000366.sql.tmp - not open
> 2011-02-25 18:07:33 CSTWARN   remoteWorkerThread_1: data copy for set 1 failed 
> 19 times - sleep 60 seconds
> 
> What am I doing wrong?

Could you give us more information about the way you settled the replication ?
Did you set the archive logging mode ?
Seems like you can't write in /tmp either bacause you don't have the right to,
or because there's no place left.

Regards,

-- 
Stéphane Schildknecht
http://www.loxodata.com