Re: Archive Dest backup with BCV
Date: Mon, 31 Mar 2008 16:30:00 -0400
Instead of doing BCV split/sync for archivelog destination you might want consider setting up so called distribution repository in terms of DataGuard for archivelog files from the primary system on your BCV box. There will be slight overhead on primary but it should be very minimal. See the following paragraph from DG admin guide for clarification:
The interim solution to this problem would be to ........ create a distribution repository on the smaller local system. A distribution repository contains only the standby control file, standby redo log files, and the standby database archived redo log files, not the data files. You would configure the primary database to send the redo data to the repository locally using the log writer process (LGWR) in synchronous mode (SYNC). Because the connection is over the LAN, the effect on performance would be minimal.
In your case ARCH could be used for log shipping to lower the overhead even more. It's kinda hybrid solution because you will be mixing DG and BCV as a media for data propagation, but it's up to you to check if that solves your business problem.
On 3/31/08, Lou Avrami <avramil_at_concentric.net> wrote:
> Hello all,
> I recently inherited several 9i databases, each a couple of terabytes,
> which are backed up using HP BCV (Business Copy).
> When a backup is executed, a script dynamically runs through all of the
> tablespaces in the database, placning them into hot backup mode. Once that
> is done, the disk sync for the datafile partitions and the archive
> destination partition is executed. After the "split" is done, another
> dynamic script is executed within the database, bringing all of the
> tablespaces out of backup mode. At the same time, the split disks are
> mounted on another server and then written to tape.
> The disk syncs/splits work fine. There is an issue with archive
> destination partition though. The sync for the archive partition takes
> several hours, mainly due to the volume of archive logs being written by the
> databases (due to other issues, which I am working to resolve). This means
> that we have to drop this mirror/split several hours before the backup, so
> that the mirror of the archive partition has sufficient time to "sync up". I
> would like to be able to keep the archive logs "online" for as long as
> possible, so that it isn't necessary to recovery them from tape if I need to
> restore a database.
> I know that the real answer is to reduce the number of archive logs being
> written. This is a political issue as much as a technical one (for example,
> I am in the process of convincing a vendor and upper management that it is
> not necessary to do daily index rebuilds!!).
> I would think though that the HP Business Copy BCV disk sync/split should
> be more efficient. Is there anyone out there who is familiar with BCV and
> how it works? If so, are there "long" delays when syncing disks? Is there
> any kind of software configuration that could help reduce the sync time? Is
> BCV the correct technology for Oracle database hot backups? Should
> something else be used?
> Any information would be greatly appreciated.
> Lou Avrami