Glad to see continuous success stories from people who were brave enough to disable chaining
Just to reassure, the chaining as a feature is not going away, but I am thinking of at least adding some nasty warning when this option is selected in the wizard, explaining the recommended way to organize required concurrency.
"Do you really want to enable chaining?" YES/NO
!_____> "Do you really really want to enable chaining?" YES/NO
!___________> "Oh gosh, you want to use it for real... ok, I've warned you"
Luca Dell'Oca Principal EMEA Cloud Architect @ Veeam Software
"Do you really want to enable chaining?" YES/NO
!_____> "Do you really really want to enable chaining?" YES/NO
!___________> "Oh gosh, you want to use it for real... ok, I've warned you"
- We use the network NBD protocol for our transport since we found we found it to be the most reliable and didn't want to access the SAN directly (only hosts have a physical SAN connection). Because of the slower speed NBD provides, and due to our desire to minimize the time a snapshot is present on a VM, we want to process our VMs one at a time. It's slower overall, but faster for any individual VM.
- We run three separate backup jobs (apps/files/emails) to reduce the size of the backup files and to make it easier to dump them to tape to send offsite.
- We process our application and file servers before our email so the email servers can be backed up during a lower utilization period.
- After doing our nightly backup we run a SureBackup job on everything to make sure our backups are happy.
- We chain our jobs so they run Apps>Files>Emails>SureBackup. This allows us to minimize the amount of time backups are running, ensure only one VM is processed at a time, and verifies our backups.