I suggest you to look into PostgreSQL log. Also you could add -x to sh. I mean:
#!/bin/sh -x of your script. This will greatly help for debugging complex scripts. You could see the out of -x at the PostgreSQL log as well. -- Tatsuo Ishii SRA OSS, Inc. Japan English: http://www.sraoss.co.jp/index_en.php Japanese: http://www.sraoss.co.jp > Thank you very much for your attention Gerd, I proved this command manually, > it work but I need to introduce a password for making the copy. > > I modified the copy_base_backup script for making a test, but I think that > this script has never been executed because the test.txt file is always > empty. > > copy_base_backup script > > #!/bin/sh > date >> /var/log/pgpool/test.txt > exit 0 > > Can you suggest me any reason or idea for resolve this problem?. > > Thank you veru much for your time. > > > ----- Mensaje original ----- > De: "Gerd Koenig" <[email protected]> > Para: "Lazaro Ruben Garcia Martinez" <[email protected]> > CC: [email protected] > Enviados: Viernes, 11 de Junio 2010 10:49:31 (GMT-0500) Auto-Detected > Asunto: Re: [Pgpool-general] Hello!! > > Hi Lazaro, > > can you perform the "scp"-command manually to check if it's working > (without entering the password) ? > > just my 50ct.....GERD...... > > On 11.06.10 16:36, Lazaro Ruben Garcia Martinez wrote: > > Jaume Sabater, thank you very much for your attention, I'm trying to make > > online recovery but I obtain a BackendError: > > In pgpool.log, I see this error but I don't know what about it is: > > > > 2010-06-08 00:17:08 DEBUG: pid 26387: I am PCP 26387 accept fd 5 > > 2010-06-08 00:17:08 DEBUG: pid 26387: pcp_child: received PCP packet type > > of service 'M' > > 2010-06-08 00:17:08 DEBUG: pid 26387: pcp_child: salt sent to the client > > 2010-06-08 00:17:08 DEBUG: pid 26387: pcp_child: received PCP packet type > > of service 'R' > > 2010-06-08 00:17:08 DEBUG: pid 26387: pcp_child: authentication OK > > 2010-06-08 00:17:08 DEBUG: pid 26387: pcp_child: received PCP packet type > > of service 'O' > > 2010-06-08 00:17:08 DEBUG: pid 26387: pcp_child: start online recovery > > 2010-06-08 00:17:08 LOG: pid 26387: starting recovering node 1 > > 2010-06-08 00:17:08 DEBUG: pid 26387: exec_checkpoint: start checkpoint > > 2010-06-08 00:17:08 DEBUG: pid 26387: exec_checkpoint: finish checkpoint > > 2010-06-08 00:17:08 LOG: pid 26387: CHECKPOINT in the 1st stage done > > 2010-06-08 00:17:08 LOG: pid 26387: starting recovery command: "SELECT > > pgpool_recovery('copy_base_backup', '192.168.101.122', > > '/var/lib/pgsql/data')" > > 2010-06-08 00:17:08 DEBUG: pid 26387: exec_recovery: start recovery > > 2010-06-08 00:17:08 ERROR: pid 26387: exec_recovery: copy_base_backup > > command failed at 1st stage > > 2010-06-08 00:17:08 DEBUG: pid 26387: exec_recovery: finish recovery > > 2010-06-08 00:17:08 DEBUG: pid 26387: pcp_child: received PCP packet type > > of service 'X' > > 2010-06-08 00:17:08 DEBUG: pid 26387: pcp_child: client disconnecting. > > close connection > > > > This is my copy_base_backup script. > > > > #!/bin/sh > > > > DATA=$1 > > RECOVERY_TARGET=$2 > > RECOVERY_DATA=$3 > > PG_HOME=/var/lib/pgsql > > > > psql -c "select pg_start_backup('pgpool-recovery')" postgres > > echo "restore_command = 'cp $PG_HOME/data/pg_xlog_archive/%f %p'"> > > $PG_HOME/data/recovery.conf > > tar -C $PG_HOME/data -zcf pgsql.tar.gz base global pg_clog pg_multixact > > pg_subtrans pg_tblspc pg_twophase pg_xlog recovery.conf > > psql -c 'select pg_stop_backup()' postgres > > scp pgsql.tar.gz $RECOVERY_TARGET:$RECOVERY_DATA > > > > These are the archiving properties in my postgresql.conf > > # - Archiving - > > > > archive_mode = on > > archive_command = '/bin/cp %p /var/lib/pgsql/data/pg_xlog_archive/%f' > > > > Can you suggest my any reason for this error??? > > > > Thank you very much for your time. > > > > > > > > > > > > > > > > ----- Mensaje original ----- > > De: "Jaume Sabater"<[email protected]> > > Para: [email protected] > > Enviados: Viernes, 11 de Junio 2010 10:12:03 (GMT-0500) Auto-Detected > > Asunto: Re: [Pgpool-general] Hello!! > > > > On Fri, Jun 11, 2010 at 4:01 PM, Lazaro Ruben Garcia Martinez > > <[email protected]> wrote: > > > > > >> Hello, I have a doubt, for making online recovery is necessary that > >> pgpool-II will be installed in all nodes? > >> Thank you very much for your time. > >> > > No. You just need one pgpool-II instance if you don't want high > > availability, and multiple pgsql nodes. Scripts executed on online > > recovery will have to be copied to all nodes, though. > > > > > > _______________________________________________ > Pgpool-general mailing list > [email protected] > http://pgfoundry.org/mailman/listinfo/pgpool-general _______________________________________________ Pgpool-general mailing list [email protected] http://pgfoundry.org/mailman/listinfo/pgpool-general
