Disaster Recovery Testing — 7
The next step with testing disaster recovery plans is rebuilding both office desktop hard disks. The office desktop is the most used computer and is the file server for the home LAN. Whereas the previous disaster recovery tests are important, this is a big test, where the proverbial rubber meets the road.
This test is not the same as replacing the disks with the backup clone disks. This test is rebuilding the disks without those clone disks. That leaves only the weekly backup disks and external sources for restoring files.
This test is unlike previous disaster recovery tests. The first gut check is that while there is a backup strategy and there are written notes, there is no formal written recovery plan or check list. I long have presumed skills and knowledge would allow meandering through any such disaster. While a reasonable presumption, a written plan or check list would be helpful. Proceeding with this test means writing a plan or check list.
Using the weekly backup disks means losing up to a week’s worth of data and file changes. Like the living room media player and Lenovo T400 laptop, using the weekly backup disks means manually recreating partition layouts on the replacement disks.
The office desktop has two internal hard disks: one 2 TB and one 1 TB. Two spare 1 TB disks are available but no spare 2 TB disks. There are additional spare disks of lesser capacity. Obtaining and storing a spare 2 TB disk might be prudent. With the additional spare disks and nominal sweat equity there is room for work-arounds.
The largest partition on the 2 TB disk is the home LAN video collection.
The immediate goal for this test is not a complete like-for-like system recovery but restoring a usable temporary system from which a like-for-like system can be rebuilt. Presumed with this disaster is that after resuming most functions and getting online that new disks would be ordered as needed.
There are three test options.
- Temporarily sacrifice the clone backup disks. This would best simulate replacing the disks like-for-like.
- Use the two spare 1 TB disks, manually create a smaller video collection partition, and temporarily forego restoring those files.
- Use additional spare disks, remap the partition scheme, and modify
I considered only the second two options. Wiping the clone disk backups was not palatable. The clone backup disks would be useful to compare file restorations from the weekly backups. Comparing would help refine the weekly backup scripts.
The video collection is important in the home network, but not having access to those files is a temporary inconvenience. Without an immediate 2 TB disk, that temporary inconvenience could be resolved with the third restoration option by using additional disks.
To conserve disk space many video files are excluded from the weekly backups because the original sources are DVDs. Theoretically, DVDs should retain data integrity for many years as long as they are stored properly. Those video ISO images would be restored from the DVDs although requiring significant time in a real disaster recovery.
There are other files excluded from the weekly backups, such as local Slackware repository mirror files.
In this specific type of disaster the hourly rsnapshot backups stored on the second internal disk will be lost. In a real disaster there is no way to avoid this. The hourly backups are intended to be a first level of defense and recovery. The hourly backups would need to start fresh after the system was restored to full use. In a real disaster of losing both internal disks the first hope is the backup clone disks remain available, but this round of testing presumes no such fortune.
Another loss is the backups of the living room media player and laptop. Those backups would need to be updated as soon as practical after restoring the office desktop disks.
So much for the head scratching and evaluations. Time to start testing.