On Sat, Dec 2, 2017 at 8:02 AM John R Pierce <pie...@hogranch.com> wrote:
> On 12/1/2017 12:44 PM, basti wrote: > > Replication is no backup. > > Its more like a RAID. > > > > That mean tubles that are delete on master by a mistake there are also > > delete on slave. > > > > correct me if i'am wrong. > You have not mentioned the version. If you are using 9.4 or above, you can apply a delay between master and slave by specifying "recovery_min_apply_delay" on slave's reccovery.conf. This might be the fastest way to recover a dropped/truncated/deleted table specially for large databases. https://www.postgresql.org/docs/9.4/static/standby-settings.html The challenge with this approach would be to choose the duration of delay. Usually something upto 6hours should be fine. If you can not identify a dropped table in 6hours, that means very likely you can go another 6 hours without it (time good enough to restore a backup and perform PITR as suggested in other answers). > > a wal archive plus occasional basebackups lets you restore to any point > in time (PITR) covered since the oldest basebackup. > > think of a base backup as a 'full' backup, and the wal logs in the > archive as incrementals. one such approach might be a weekly > basebackup, where you keep the last 4 weeks, and keep all wal files > since the start of oldest basebackup. yes, this will take quite a bit > of space > > > -- > john r pierce, recycling bits in santa cruz > > > -- -- Best Regards, *Sameer Kumar | Senior Solution Architect* *ASHNIK PTE. LTD.* 36 Robinson Road, #14-04 City House, Singapore 068877 T: +65 6438 3504 | www.ashnik.com Skype: sameer.ashnik | M: +65 8110 0350
