BCV(Business Continuity Volume)

EMC for DBA's

Symmetrix refers to the EMC disk devices, disks within thid symmetix can be divided into various sized logical volumes, these logical volumes can be logicaly joined together and presented to a server.

There are 2 types of logical volume:

Hypervolumes - range of continuos space on a single disk. A single disk can be divided into 128 parts or hypervolumes this is refered to as slicing or splitting. These now form the basis to create metavolumes.

Metavolumes - A group of hypervolumes of the same size grouped together, this is then presented to a server. Once part of a metavolume, only the whole meta can be presented to the server not single disks.

Meta Head and Tail

When hypervolumes are added to a metavolume they are given a sequence, the first hyper being the head and last being the tail. Data is written to Meta from head to tail.

Concatenated Metavolume

Writes data to one hypervolume at time until it becomes full, then moves onto the next

Striped Metavolume

Data is written to metavolume in multiples of cylinder size. The first write will go to first meta (head), then the next write to the next meta and so on. Amount of data written to the hypervolume depends on the stripe size. The stripe size is based on being a minimum of 2 cylinders. A cylinder is 480K so minimium stripe size is 960K.

Server disks in EMC are usually refered to as R1 disks, copies of these disks will be known as R1 BCV disks. Remote disks kept in sync with SRDF would be called R2 disks.

BCV

A BCV (Business Continuity Volume) is a mirrored on demand set of disks. This is usually used for backup purposes to capture a copy of a database. With a cold BCV backup the Oracle database would be shutdown first. With a hot backup 2 disk volumes assocated with the BCV would be defined. The database would be put in backup mode the a snapshot of disk volume 1 would be taken that contained all the datafiles. After this a log switch would occur then a snap shot of the second disk volume that contained all the archive logs would be taken.

SRDF

This is a remote in sync/async copy of a servers disk, usually used for DR purposes, so if a production site was lost, the data would be upto data at another site (The emc version of dataguard/standby)