Ühel kenal päeval, R, 2007-09-07 kell 12:03, kirjutas apoc9009:
> Andrew Sullivan schrieb:
> > It seems that what you want is near-real-time online backups with _no
> > cost_, which is not a feature that I think anyone will ever work on.
> > A
> 100% Correct!
> I think anyone commit the Statement, thadt a Databases is a very
> imported Part of Software
> for a wide range of Peoples. It is absoluty normaly thadt thoose Peoples
> who using Databases having the Wish to have
> there Data save and up to Date until to the Last Record of a Customer
> (for Example: A Order form a Onlineshop).
> Very Large Databases with sizes more then over 1 Terrabyte cannot be
> backed up continously using Dumps
> or the common Backupmethods
> The Abstract need is:
> 1.An exact Copy (1:1) of the Productiondatabase should be mirrored at a
> diffrent Machine, connected via (LAN/WAN) (as Backup)
> 2.The Backup Operations should be done normaly while the Database is
> Online with very small Bandwich, I/O and CPU Trafficcosts.
> 3.Per Month or Week (or choosable) at a New Backup-Timeline/Version
> should be started (depend on the Basebackup)
Probably you don't want your backup as a base backup + a bunch of
If you need to rally move to backup, then even a weeks worth of logfiles
will take several hours or possibly days to apply.
What you want is a standby machine, which applies all WAL files as they
are copied over.
> 4.A New Base and its Streaming Backup Configuration should be
> "Hot-Addable" to an running Productiondatabase (without Shutdown or Lock)
It currently is
> 5.A Remote Operationmanager Interface (GUI) should be there for
> Monitoring and Maintaining the Backups (maybe in PGAdmin)
Huh ? I don't think this has anything to do with postgres-hackers list.
Actually your other questions should also go to some users/newbies list,
as this is a question of using existing features, and not something that
needs to be added to backend.
> 6.If the Production and Mirror replicated Database is crashed (Hardware
> Failure or Provider Problems) the Recovery should done verry fast.
Yep. That is what Hot-standby as described in postgres manual is meant
> (An Idea for 5. / Clientside)...
> It would be a great Feature, if the PSQL-Client having a Local Deamon,
> thadt can Download the Backup Catalogues from
> the WAN-Backup FTP-Server continously down to the local Adminmachine.
> Then the Admin is able to burn the DB-Backups on
> DVD or write it on Tape on its local Machine at his Company (if the
> Backupreplicationserver fails two or the Internetprovider has
> Problems, the DB-Admin can apply its local Backups from DVD to the New
> DB on a New Machine anytime)
All this keeping the WAL logs on CD's/tapes has nothing to do with fast
recovery after failure.
This is something that would be useful, but for entirely other purposes.
Namely for a case, when you need to _GO_BACK_ to an earlier state. For
example you discover that an careless operator or software bug has
deleted important data last Wednesday and you need a way to get it back.
In that case you take your last Sundays base backup and apply WAL up to
few moments before the error happened. But it will not be fast.
---------------------------(end of broadcast)---------------------------
TIP 7: You can help support the PostgreSQL project by donating at