[Pgpool-general] [pgpool-II] PITR online-recovery

[Anton Koldaev]

I'm using pgpool-II + PostgreSQL 8.4 on three servers: primary + standby1 +
standby2. Replication mode is "on" Load balance mode is "on" (between
primary and standby1)

I followed the official tutorial to configure online recovery with PITR:
http://pgpool.projects.postgresql.org/pgpool-II/doc/pgpool-en.html#online-recovery

There is a tar command in script "copy-base-backup":

 tar -C /data -zcf pgsql.tar.gz pgsql

So all my PG cluster directory being copied while script is running for a
very long time:

ls -1 /srv/pg/data/
PG_VERSION
archive # directory with postgres archive files 100Gb+
backup_label.old
base
global
pg_clog
pg_hba.conf
pg_ident.conf
pg_log
pg_multixact
pg_stat_tmp
pg_subtrans
pg_tblspc
pg_twophase
pg_xlog
pgpool_recovery
pgpool_recovery_pitr
pgpool_remote_start
postgresql.conf
postmaster.opts
postmaster.pid
recovery.conf
recovery.done


   1. Is it neccessary to copy all of them from primary node? Where to find
   the correct list of what do I need to tar?
   2. What directories can I clean up on the standby node when I need to
   recover it after failover?
   3. How to safely remove old archive logs from primary and from standby
   nodes? There are about 30Gb of archives per day on my nodes :(


-- 
Best regards,
Koldaev Anton
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://pgfoundry.org/pipermail/pgpool-general/attachments/20110616/a0a9c830/attachment.html>