Hi All, even with PITR not in place yet, how far is it possible to use the xlog to restore a database (beyond redo-after-crash)? Suppose I lose my data partition, but have an archive of my xlog and clog, and occasional pg_dumps, it probably won't be as simple as restoring from the dump, stopping pg, copying in the archived xlog and clog, and restarting the pg (to get an effect similar to playing back the redo logs in oracle) ... I am not sure whether the whole question makes sense, so please tell me if it doesn't. I am trying to get a backup that is more fine-grained than just snapshots. (Or, with the MVCC model and database sizes not more than a GBor two, are pg_dumps so cheap that I can do them much more frequently than a full backup in oracle?) My search for backup solutions on the web gave me either wrappers around pg_dump, or something in the direction of replication and mirroring, many of which seemed out-of-date ... can anyone point me to the "preferred" backup solutions out there? Thank you for your time (in light of my slightly confused questions), Colin ---------------------------(end of broadcast)--------------------------- TIP 1: subscribe and unsubscribe commands go to majordomo@postgresql.org
