| From: | "Dann Corbit" <DCorbit(at)connx(dot)com> |
|---|---|
| To: | "Bruce Momjian" <pgman(at)candle(dot)pha(dot)pa(dot)us>, "Lamar Owen" <lamar(dot)owen(at)wgcr(dot)org> |
| Cc: | "Hannu Krosing" <hannu(at)tm(dot)ee>, "Jan Wieck" <JanWieck(at)Yahoo(dot)com>, "Christopher Kings-Lynne" <chriskl(at)familyhealth(dot)com(dot)au>, "HACKERS" <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: (A) native Windows port |
| Date: | 2002-07-05 22:30:27 |
| Message-ID: | D90A5A6C612A39408103E6ECDD77B82920D046@voyager.corporate.connx.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On generic recovery...
What is wrong with this strategy...
0. Put the database in single user mode.
1. Dump the Schema, with creation order properly defined, and with all
constraints written to a separate file. (IOW, one file contains the
bare tables with no index, constraint or trigger stuff, and the other
contains all the RI stuff.)
2. Dump the tables (one by one) to text files with "copy"
3. Create a new database in a new location.
4. Feed it the bare table schema
5. Pump in the table data using "copy" from the saved text files
6. Run the RI script to rebuild index, trigger, PKey, FKey, etc.
I find that is the most trouble free way to do it with most DBMS
systems.
Am attempted dump from DBMS X.Y and a load to DBMS (X+1).Y is always a
pile of trouble waiting to happen -- no matter what the system is.
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Sander Steffann | 2002-07-05 22:44:07 | Re: I am being interviewed by OReilly |
| Previous Message | Alvaro Herrera | 2002-07-05 22:14:28 | Re: I am being interviewed by OReilly |