Discussions related to exporting backups to tape and backing up directly to tape.
Post Reply
cdlane
Enthusiast
Posts: 53
Liked: 3 times
Joined: Aug 09, 2015 5:48 pm
Full Name: Chris Lane
Contact:

Revisiting our Setup

Post by cdlane »

Hi All,

We have been running Veeam for a couple of years now. When we started, we had an old DL380 G7 running Linux with limited disk space and used this as the repository with Reverse Incremental backups.

A few months ago we purchased a Synology DS1817+ with 22TB of disks. This is linked to our VM Host via a dedicated 10Gbps card and uses iSCSI to present a mapped drive on the host which Veeam backup server then throws data at it once a day with a stupidly large retention. This may or may no be the best way to do it but it works (a bit slow despite the 10Gbps card - the disks are the issue I think). Any suggestions/improvements welcome.

After the backup has run a Cloud Connect job copies the data to our Veeam Partner. We then run a Sure Backup job but not pertinent to this post.

With an apparent resurgence in tape et al I have been mulling over the "Cryptoware" question; are we safe; does Cloud Connect offer a suitably big air gap; should I consider AWS Tape Gateway + Glacier; Am I overly paranoid?

According to the Veeam 321 model we have 2 copies of data in one format with one copy offsite.

So my questions are: -

i) AWS and costs, can anybody give me a steer on how to price this? Several articles talk about Veeam --> Amazon S3 --> Amazon Glacier but then AWS's calculator talks about "PUT/COPY/POST/LIST Requests" etc...how on earth can I calculate that ahead of time? Can I go direct to Glacier?

ii) Forever Forward/Forward Incremental/Reverse Incremental - Any pro's and con's re these when copying to tape? I assume the current Reverse Incremental would basically mean a whole backup per day sent to Amazon and so (I assume) would cost a pretty penny.

iii) GFS, Hanoi etc thoughts?

Sorry if some of these seem obvious, I'm trying to move to something more elegant than just throwing data at a disk with one daily job.
Dima P.
Product Manager
Posts: 14417
Liked: 1576 times
Joined: Feb 04, 2013 2:07 pm
Full Name: Dmitry Popov
Location: Prague
Contact:

Re: Revisiting our Setup

Post by Dima P. »

Hello Chris.
Am I overly paranoid?
Not at all. We must not underestimate cryptolockers threats and offsite or offline copy sounds like a perfect protection, so tape backup is a decent option.
AWS and costs, can anybody give me a steer on how to price this? Several articles talk about Veeam --> Amazon S3 --> Amazon Glacier but then AWS's calculator talks about "PUT/COPY/POST/LIST Requests" etc...how on earth can I calculate that ahead of time? Can I go direct to Glacier?
If we are talking about AWS Gateway VTL (sending data as emulated tape media), you should check the gateway subscription fee plus storage fee.
Forever Forward/Forward Incremental/Reverse Incremental - Any pro's and con's re these when copying to tape? I assume the current Reverse Incremental would basically mean a whole backup per day sent to Amazon and so (I assume) would cost a pretty penny.
Well. I'd say you can use any backup mode that fits your primary backup model. Tape jobs can work with any type of backup files and send them to tape (for incremental we even have synthesized full backup option which allows to create a full restore point from increment on disk to split up the long incremental backup chain with full).
GFS, Hanoi etc thoughts?
GFS is the only option nativly integrated in Veeam B&R console as GFS media pool for tape media.
cdlane
Enthusiast
Posts: 53
Liked: 3 times
Joined: Aug 09, 2015 5:48 pm
Full Name: Chris Lane
Contact:

Re: Revisiting our Setup

Post by cdlane »

Hi Dima,

Thank you for your response, nice to know I on the right track.

I was interested in your comment "use any backup mode that fits your primary backup model" as I use what was put in place by a consultant when it was introduced (Reverse Incremental) and in true "If it ain't broke, don't fix it" style, I have never changed it so some research is needed there - the pro's and con's and what we need etc.

I have the AWS VTL up and running and it was very promising for the first half our, crunching along at 84Mbps...then it dropped to about 8-11Mbps and I canned the job as it was going to take days to send the data to AWS as oppossed to hours. Still haven't worked out whats happening there. I have read some best practices etc but we don't have the luxury of separate spindles and SSD's etc. We have one Hyper-V host with a set amount of memory and disk and one 100Mbps fibre lease line and that's it.

On the basis of the above I might have to call this one a bust which is a shame.
Post Reply

Who is online

Users browsing this forum: No registered users and 31 guests