We use a three pronged approach at data storage. We have a NAS/SAN device that takes snapshots of all data 4 times a day, 5 days a week, retaining 6 weeks at a time. We also replicate those snapshots across to a mirrored SAN/NAS device at our secondary datacenter. We then use backup tapes to backup the secondary NAS/SAN device as well as the system partitions and “sytem states” of our servers.
Our reasoning is that if data gets corrupted, then it also is corrupted on the snapshot. If it is not frequently accessed data, then that corruption may go unnoticed for a while. If our snapshots are all corrupted, then we will need to restore from backup tapes. Also, should something totally off-the-wall occur to both SAN/NAS devices, we still have fairly current data to restore from.
With virtual technology, we will no longer have to run backups of system partitions and “system states”, as that data will reside within a virtual session. Our NAS/SAN also has the capability to snapshot virtual sessions while they are running, so those virtual sessions will also be snapshoted 4 times a day, 5 days a week, for 6 weeks and also mirrored to our secondary site. We will then use backup tapes to backup those virtual sessions along with the respective data for the same reasoning as stated above.
Brad G.
I can say for our company it’s all over the board. Some people are still using legacy tapes and some are on NAS (whether inhouse or 3rd party).
Tapes are very expensive and not as scalable but NAS, despite this economy of abudance, still isn’t cheap…but it is getting cheaper.
Jon D.
We use a three pronged approach at data storage. We have a NAS/SAN device that takes snapshots of all data 4 times a day, 5 days a week, retaining 6 weeks at a time. We also replicate those snapshots across to a mirrored SAN/NAS device at our secondary datacenter. We then use backup tapes to backup the secondary NAS/SAN device as well as the system partitions and “sytem states” of our servers.
Our reasoning is that if data gets corrupted, then it also is corrupted on the snapshot. If it is not frequently accessed data, then that corruption may go unnoticed for a while. If our snapshots are all corrupted, then we will need to restore from backup tapes. Also, should something totally off-the-wall occur to both SAN/NAS devices, we still have fairly current data to restore from.
With virtual technology, we will no longer have to run backups of system partitions and “system states”, as that data will reside within a virtual session. Our NAS/SAN also has the capability to snapshot virtual sessions while they are running, so those virtual sessions will also be snapshoted 4 times a day, 5 days a week, for 6 weeks and also mirrored to our secondary site. We will then use backup tapes to backup those virtual sessions along with the respective data for the same reasoning as stated above.
Rusty H.
Jon, sounds like you guys have a great plan. Thanks for sharing.