We have Veeam replication every 6 hours which is our RPO, basically if we create file and then delete it in this time frame we wont recover it. How would we lower our RPO to lets say 30 minutes ? Run replication with 30 minutes intervals ?
It takes up to 30 min just to finish a single vmware machine so there must be different way ? What about VSS, I have never used that feature - is that what i'm looking for here ?
Unfortunately, I do not understand what different way you're talking about. What time does it take you to perform an incremental run? 30 min or it is for a full job run? By the way what is your replication target?
As for the VSS option, then please look through our sticky F.A.Q. to know more about application-aware image processing.
I think he means that file is created and deleted within 30 minutes, its lost because it wasn't backed up.
By VSS, he may be referring to the Shadow Copies feature in Windows Server (2003+), which is a pretty good option for Windows boxes. Be sure to set space restrictions as if you're taking shadow copies every 30 minutes, I'm sure it will add up quick..
Yes, there should be two ways of recovering "30 minutes" files, either use Shadow Copies and perform File-Level restores if needed or create replication jobs with 30 minutes interval.
Anyway, I still wonder what time does incremental replication job pass take and what is the destination target for these jobs.
Creating a replication interval of 30 minutes WHEN the replication job itsself takes about 30 minutes would be no best practice at all The next following Replica would fail because a snapshot would still be active in use.
best regards,
Joerg
Hi Anton - ofcourse, a reported failure would not occur. The rep cycle would be skipped after it gets aware of the active snapshot.
best regards,
Joerg
Veeam replication jobs can be scheduled by the minutes, so change your schedule to 45 mins for example to lower your RPO. You could also choose the Continuous schedule option. That way you won't have to Guestimate how long your replica cycle will take, whether a full or incremental.
Of course, a constant replication job means a snapshot is constantly being created and committed every 30 to 45 mins. Hopefully your mix of VMs on the same datastore can co exist with this decision?