Direkt zum Inhalt

Incrementals an accident waiting to happen?

Thread needs solution

I was starting to play with backing up to an ftp site but to start out wasn't actually using an ftp site. To use and ftp site you almost have to use incremtals because of the speed of the internet so here is where I started playing: (using BR10 workstation)

I did a full disk image of my computer and copied the resulting tib to a location on my file server. I then set up another backup plan to do an incremental of my computer to the same location I copied the full image. Of course my thought was to see if this incremental would append to the full. Of course it did not work. The incremental did nothing more than make a second full to that location.

I read a couple post here about an incremental doing not much more than making a new full but the explanations I don't think were completely accurate. The senario was making a full with WinPE and than trying to do incrementals from Windows. The explanation was that WinPE and Windows were using different users to create the backups thus the incremental really made just a second full.

That might be the case but via my senario I think there is a lot more comming in to play than just the user creating the full and incrementas being different. It looks almost like the first full and all subsequent incrementals must be made from the same plan. It also looks like the meta folder and job ID somehow come in to play.

Back to the FTP backups... doing a full to an FTP is pretty much out of the question because of internet speed. How would you "seed" (make a full of the server to something like a USB HD and copy it to the FTP site at the remote) and then write a plan to do incrementals to that FTP site that will recogize the full I copied there and append to it?

Exactly what does a incremental look for or look at to determine if a full already exists and it is OK to append to it?

0 Users found this helpful

1) Did you copy .meta as well?
2) Stupid question - did you check that the second slice was created as 'full' by browsing vault? If not, could it happend that you left the original full .tib on your disk so that the second backup plan backed this tib in the incremental mode resulting in the large incremental slice?
ps - I tried to reproduce it using the second vault on another place on local disk and incremental of reasonable size was created as expected.

It is interesting :) I found another thread here that explained how to "seed" a vault with a full image and indeed when I wrote an incremental it did actually do an incremental. That meta folder is all critical to incrementals and differentials :)

Proceeding to the next phase is there any way you know of to at this point "import" this new vault in to AMS? Perhaps a better explanation. I want to backup to an FTP and this FTP will be hosting AMS so AMS can do the validating and consolidating and take the burden off the remote. I guess what I am really trying to do is manage a "machine" from a remote AMS. Can you guide me in the right direction?

The real issue I am having at this point is that I can't seem to find where to add a consolidation task soley and by itself. No backup, no retention just a straight consolidation task.

Convert it in-place a unmanaged vault? Maybe just creating a new unmanaged vault and then copying archives and .meta into the vault folder will do it. In case of managed vault, probably the only way is to export, and it will load the network unless the agent performing export is installed on the same machine as the AMS.

Getting closer... :)