Hello,
We have built a database application build around PostgreSQL v7.03.
And we're very happy with PostgreSQL functionality & performance.
In a live situation 50-100 clients are inserting and updating database
records. It is very important that we have all the data available on
the back-up servers in case the main server fails.
After looking briefly at the the RServ 0.1 replication software we decided
it was not an option (for now) (it had serious problems in our tests,
there is no documentation available and inquiries about it to PosgreSQL,
Inc. remain unanswered unfortunately).
So we decided to use a compromise and do frequent pg_dump runs from the
backup machines to the main server. It all seems to work fine.
But I still have some worries: is it OK to run pg_dump on *live* databases ?
Do you get consistent dumps from it when there is inserting and updating
going on ?
In the v7.0.3 documentation I can't find any info on this issue.
But in the v7.1 docu:
http://www.de.postgresql.org/devel-corner/docs/admin/backup.html
I read:
"Dumps created by pg_dump are internally consistent, that is, updates to the database while pg_dump is running will
notbe in the dump."
That looks great. But we are using v7.03. Does this mean that the v7.1
documentation is just more detailed and that the same applies to v7.03 ?
Or is this a feature that is available only with v7.1 but not with v7.03 ?
Any help is much appreciated !
Friendly greetings, Rob van Nieuwkerk