Configuring your Environment for Software RAID
Configuring the Package Control Script and RAID Configuration File
| For example, let us assume that the data storage links in Figure |
| fail before the heartbeat links fail. In this case, after the time |
| specified by Link_Down_Timeout has elapsed, a package in |
| Datacenter1 (DC1) will continue updating the local storage, but not |
| the mirrored data in datacenter2 (DC2). While the communication |
| links must be designed to prevent this situation as far as possible, |
| this scenario could occur and may last for a while before one of the |
| sites fails. |
|
|
NOTE | For more information on how access to disks is disabled in certain |
| failure scenarios, see “Setting the Value of the Link Down Timeout |
| Parameter” on page 69. |
| Let us consider a few failure scenarios and the impact of different |
| |
| RPO_TARGET values. The discussion below is based on the timeline |
| and events shown in Figure |
Figure | RPO Target Definitions |
To ensure that no data is lost when a package fails over to DC2 and starts with only DC2's local storage, t2 must occur between t0 and t1.
80 | Chapter 3 |
