Hi All,
I have not used backup to disk in BE12 before, we are just going straight to tape.. but due to the type of data we have (millions of small files and directories) this cant be read fast enough into the tape drive, and this causes abysmally slow backup speeds. I understand this is because the tape drive writes some data, then has to wait for more data to come from the disks, during which time it has to rewind the tape a bit and start writing again.. and repeats this process many times.
I have found that backup to disk seems to handle this data much faster because the target disks can write as slow or fast as the source can provide them. The tape unit can then backup the massive single backup file onto tape much faster since its a single file on disks that do nothing but backups (this was just testing with NT Backup but I assume that BE12 will offer roughly the same performance gains).
There is still a lot of data, so we have to back up some folders as separate jobs.
I would like to have a single B2D location which is overwritten after it is duplicated to tape by the next job if that is possible.
A backup job runs on Folder1 - this is a backup to disk job
A duplicate job runs to duplicate this disk job to tape
then..
A backup job runs on Folder2 - this is a backup to disk job that overwrites the last B2D job output
A duplicate job runs to duplicate this disk job to tape
etc..
Since the seperate volume I am going to use for B2D will not be large enough to hold all of the data I want to backup, I assume that as long as its big enough to backup any single folder that my idea will work.
Does this sound like a possible solution, or am I going about this the wrong way?
Thanks!