Assure MIMIX for AIX records write information into log files. Logs are block storage devices of the same size. Typically, multiple logs are contained in a Logical Volume. The pool of disk storage that contains the logs is called a journal. A production journal exists on the production server and a recovery journal exists on the recovery server.
The production journal is the storage that contains all of the logs. A single log is transferred to the recovery server when that log is filled. For example, if each LFC is 64MB and there are 100 production LFCs, then the production journal is 6400MB. When the current LFC is filled with approximately 64MB of write I/O data (there is some additional metadata), it will be transferred to the recovery server.
All logs in the production journal are redologs. They contain information that moves the disk image of the application forward through time when the information is applied. This is called rolling forward.
Half of the logs in the recovery journal are redo logs, and half are undologs. Undo logs contain information that moves a disk image of the application back through time when they are applied. This is called rolling back.
The recovery server also contains the snapshot journal. The snapshot journal is the space on the recovery server where Assure MIMIX for AIX stores copy-on-write information and write-cache data for snapshots.
The following table shows the variables that are used for estimating journal sizes and log sizes. You need these estimates in order to configure Assure MIMIX for AIX:
|
Concept |
Meaning |
|---|---|
|
Throughput |
Disk write rate of the application over a business cycle, typically 24 hours. |
|
Average throughput |
Average write rate of the application during a business cycle |
|
Peak throughput |
Maximum write rate of the application during a business cycle. |
|
Peak duration |
Average length of time that peak throughput occurs. |
|
Bandwidth |
Maximum throughput supported by Assure MIMIX for AIX. It is typically limited by the network that connects the production and recovery servers. |