| From: | "Mag Gam" <magawake(at)gmail(dot)com> |
|---|---|
| To: | pgsql-novice(at)postgresql(dot)org |
| Subject: | On-line backup and point-in-time recovery (PITR), and human error |
| Date: | 2007-11-22 16:41:23 |
| Message-ID: | 1cbd6f830711220841n2f7457c2x1990df55d8af71e3@mail.gmail.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-novice |
Hi All,
I am currently building a database where we get batch feeds from various
vendors. On a daily basis we generate about 5000 updates. Sometimes the
vendor may send us a bad feed/data, and we have to stop all processing,
restore a previous "good" database, and ask all vendors to resubmit their
jobs.
Our current backup strategy is, we do a pg_dump of the database on a nightly
basis. Each vendor sends their feed on a hourly/by minute basis. The vendor
will inform us, when they send a "bad" feed. Its becoming tedious for us to
restore, and have all vendors resubmit their jobs.
My question is, would PITR help us in this situation? If a vendor states
they sent a bad feed at a certain time, can we just stop all transactions
and roll back our logs, and proceed from there? Any thoughts for this
problem?
TIA
| From | Date | Subject | |
|---|---|---|---|
| Next Message | chrisj | 2007-11-23 19:13:50 | Re: [OT] Installing 8.3beta on debian (stable) |
| Previous Message | Sean Davis | 2007-11-22 13:37:20 | Re: xpath question |