Hi, i'm currently running a tandberg tape library. I've set up media pools for 3 different types of jobs, one is a full backup each night of one server, the others are incremental jobs with a weekly full back up of the remaining servers. I have set the overwrite protection on the full jobs for 14 days, so they should never be overwritten once they have been used. I do however remove the full tapes from the library, once they have been run.
My question is though, the incremental jobs run each night, and only take up about 25-30 gig, the vxa tapes can hold obviously a lot more than this, so i have setup the media pool for these jobs with a overwrite protection period of 10 days, but they are appendable. My take on appendable would be that each previous job would not be overwritten on the tape until 11 days after the backup job was run, BUT my take on appendable is that other future backup jobs would be written to another section of the tape, and then be protected from being overwritten.
Currently if i put a tape in with overwrite protection enabled, even though the tape is appendable, the job will not start until a overwritable tape is put in the library.
is it me or is something not right here?
OPP (overwrite protection period) is for the whole tape, and does not protect individual areas of the tape. So right now, you have told your system to protect the tape from ANY writing for 10 days, so it is asking for an overwritable tape.
In reality, since you are using the same tape daily, what you want is no OPP (setting 0), but an append period of 10 days, if I understand your needs. By setting the append period to 10 days, it will append for 10 days (unless you run a job on that tape that specifically tells it to overwrite), and then on that 11th day, it will start again and overwrite it.
In practice, here is an example. I have an LTO-3 tape (400/800GB) that backs up 87GB daily. The OPP is 0 and the Append period is 6 days, so I always only keep one week's worth of data on it at any one time, which is what I want.
many thanks for the reply. I think you may have hit the naqil on the head. I was thinking that BUE was going to work in a similar fashion to other backup applications, and could multi session a tape.
I have one question on your suggestion though,for the weekday incremental jobs, would i only need one tape, and that a full weeks worth of "appended" incremental data would be written to the one tape?
You never specified how big your tapes are, only that each job was 30GB tops. I was assuming that since you have a library, that maybe your tapes were big enough to store 120GB on one tape, and in that case, Mon-Thurs would easily fit on one tape. You just have to do the math.
If not, you can use two tapes for the incrementals and swap tapes on Wednesday morning, or allocate a second slot, one to Mon-Tues, and the second to Wed-Thurs. I don't have enough details to say exactly how I would do it, but you get the idea.
Lastly, if your tapes do hold a lot of data, you should run an differential right before your full backup and see if your tapes can hold a differential. Doing nightly differentials would make restoration that much easier if the need arises since you only would have to restore two tapes maximum. Right now, you have to restore your full and every incremental in order.
the tapes are the max the library can take, and will go up to 320 Gb according to tandberg.
The data being backed up at the moment across all of the servers is about 205 Gb, and each incremental backup is about 9 Gb, so there is plenty of space on the tapes.
Point taken about the differential backups, i was origiannly advised to use incrementals, but as i've not had to call upon the backups yet, i've not had to carry out a restore.
Saying that, once the plan is sorted and in operation, i'll start to carry out a few test restores.
Thanks for your help / advice Kevin
I have one piece of advice for every network administrator: Practice your disaster recovery of your most important servers before you need it. Although you need to be careful, make an image of your server, replace the disks with empty disks, and try restoring. Then replace the original disks if it does not work.
If you are uncomfortable doing this, practice with a workstation first.
Most people test the restore of individual files, by necessity, but neglect the disaster.
I hear what you're saying, and for that reason love Acronis. It's got me out of the mire a few times.
The reason for my post was down to the fact that up to recently, i've been able to take a complete backup of everything each night. Due to an increase in data the complete backup went out of the window and i had to start using incrementals.
Many thanks for you help & advice.