Hiya David, I'm assured by our DBAs that this one-tablespace-at-a-time method works - it's apparently what Oracle recommend, and they do restores and recovers frequently for things like creating test databases, so it's a proven method. It could be that my description of the process was rather lacking, there's likely other details around it, but the basic theory is sound. Hot-backup one tablespace at a time, followed by a log switch and back up all the logs.
Using RMAN is one of the options I'm looking at, but I haven't managed to spot a good description of how to integrate it into Bacula to achieve a backup straight to tape, either a directly attached drive or across the network. Ideally, as I think I mentioned previously, I'd like to configure our SAN fabric to give one of the drives (in our autochanger tape library) directly to the database server, and back up straight to tape without going across the network. But it's not at all clear to me how to do that with Bacula and still be able to use Bacula and the autochanger to switch tapes and keep track of volumes. Cheers, DJ. -----Original Message----- From: David Ballester [mailto:[EMAIL PROTECTED] Sent: Wednesday, 26 November 2008 08:02 To: David Jurke; bacula-users@lists.sourceforge.net Subject: Re: [Bacula-users] How to set up large database backup El dt 25 de 11 del 2008 a les 13:12 -0500, en/na Dan Langille va escriure: > On Nov 25, 2008, at 2:10 AM, David Jurke wrote: > > > The problem I have is with our large (expected to grow to several > > terabytes) database server. I'm told by the DBAs that the size and > > amount of activity on this database is such that putting the whole > > database into hot backup mode for the several hours it takes to back > > it up is a Bad Idea, it generates far too many log(?) files. The > > method they recommend is to put a single tablespace into backup > > mode, back that up, put it back to normal, repeat for each > > tablespace. The backup takes the same time, but doesn't generate > > anything like the amount of log(?) files. > > There is no way to dump interactively? I'm a PostgreSQL fan and > creating a backup doesn't add overhead. > > Are the DBAs sure that this won't create an inconsistent backup? That > is, to restore, you are combining data from different times. I'm sure > that transactions have occurred during that time... etc. Backing up only Oracle datafiles ( tablespaces ) making up 'backup mode' doesn't save you to backup an INCONSISTENT datafile. In backup mode, the datafile blocks are modified AS USUAL, but the datafile header is freezed, when turning on the tablespace 'bakup mode' off, the smon process executes a recover of this datafile using the redo/archive info. To get a recoverable database, you must put ALL tablespaces 'backup mode' on, copy the archives, put them again on backup mode off and force archiving of the current logfile, and copy all archives generated between the tablespace modification and 'now'. OF course after that you must backup a copy of control file, init parameters, etc... USE RMAN rman in 9i is able to know where is the tablespace highwatermark and only backup the data before it ( with your method, you copy all datafile, even it is empty of usable data ), it does compression on the backup result. rman in 10g is able to ( rman 9i + ) know the unused blocks under the tablespace highwatermark, backing up only the used blocks My 2 cents Feel free to give my mail to your dbas, I'm always open to help ( I'm usually stay on #oracle @ irc.freenode.net ) D. ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
