[POLL] How do you take Veeam backups off-site?

[Gostev]

The results of this poll will be available publicly, however the voting itself is anonymous (the forum does not keep track of who voted for which option).

[mongie]

I have two different methods - but can only vote once... so you can add one to the "TwinStrata" category.

[Wintel]

what constitutes 'off-site'?

We have three sites: two sites host LOB servers (vmware and cross-site storage replication) and the third hosts the backup repository. they are all separated by ~1km.

Is this considered adequate for off-site? we don't currently do any further copy to tape or fourth site

[Gostev]

There is really no universal definition of "off-site". It is whatever distance you feel is sufficient to protect against possible disasters... and mostly depends on your physical location, and associated potential threats.

1 km is good enough protection from plane crash. 1 km might also be sufficient protection against fire in cities and rural areas, but not in deserted and hardly accessible areas (I've experienced some wild fires in Wyoming in 2002 first hand, and those can go crazy very fast). Likewise, 1 km was clearly not enough for New Orleans during Katrina, or during tsunami in Thailand in 2004. And, of course, 1 km will not be enough during civil disorder or war (and I know of at least one large customer of ours for whom the latter threats are very real at the some sites).

But then again, since certain large-scale disasters will destroy some types of business completely anyway, you may decide not to protect the data against them either, and this would be perfectly valid approach...

Although this goes both ways in terms of scale. And people seem to overlook even most basic stuff. I'd really like everyone who are not keeping copies of their backups to at least start thinking about protection from burglars, upset employees and other types of criminal activities. Even simply storing a copy of your backups in an adjacent room might be good enough "off-site" backup strategy to handle this specific type of threats. Somebody with a hammer could be in your server room right now...

[dstacbsu]

We have 5 physical sites connected by fiber (some 2+ miles away.) One site is the "hub" and has the vmware servers. A second site has the VEEAM backup server that backs up to a QNAP NAS(also at the other site)... the QNAP VEEAM storage is synced (SyncToy) to an external USB HD after every backup run... We plan on cycling out 3 different USB HDs and storing the unused HDs at a third site, but we have not started doing this yet.

[TommyTheKid]

We were doing the "copy the backup files to a USB disk and take it offsite" after a couple failed attempts at an offsite Veeam job, but now we have pared down the number of VMs (critical only) going offsite and upgraded the Internet connection so that the offsite backup is able to complete in about 12 hours. It has been nice to just know they are offsite. It would be nice to maybe push them to the cloud too but I just don't see the cloud as a viable *recovery* option. How long would it take to download a full backup, and any requisite incremental backups from the cloud? (~3TB in our case) I also haven't found a reasonable way to handle offsites over WAN other than a separate Veeam job with a remote WINDOWS repository (vm). The Linux REPO doesn't appear to have the brain power necessary to recover from a "failed" backup without reading the entire file and sending it back over the wire to the veeam server for processing (guh). Sending to a remote SMB share is just a problem waiting to happen (over the Internet especially). If you just sync the files, you end up syncing the full, often, and sending a full backup over the WAN would take a week or so (at least in our previous attempts). Anyhow, long story short is we are using a separate remote Veeam job

~tommy

[Trev]

We use Veeam to backup to a lower cost 16TB Drobo SAN for online storage and then use Veeam to copy to removable disk for offsite recovery if needed.

[mcwill]

We don't maintain copies of backups offsite, instead we Replicate to our DR site and maintain archival Backups onsite.

[andybh]

We consider backup and DR (or even BC) as two separate issues. For Business continuity we have remotely hosted replicas that are kept up to date with DoubleTake CDP. Backups are kept onsite currently but will be interested to look at how Veeam can make use of Tape Libraries from Version 7.

[thuizenga]

I have two different methods - but can only vote once... so you can add one to the "TwinStrata" category.

Same here, since Veeam has no way to make secondary copies we are performing a backup to local disk and performing a secondary veeam replication job offsite.

[ChrisStarmer]

A couple of things that I warn my customers to protect themselves against are "security lapse" and "disgruntled employee".

If you have all your backups "online" - on NAS/SAN drives locally or in a DR center offiste then these are still vulnerable to someone infiltrating your network, deleting all your backups and live machines. Tapes or removable disks protect against this.

More likely is the "disgruntled employee" attack. You have fantastic DR procedures and someone takes tapes offsite - but if the same person has all the passwords and the tapes then you'd better make sure you don't upset them !

I warn customers against this, however few resolve the issue.

Just some thoughts.......

[dellock6]

In security terms that's called segregation of duties, that is not having a sungle people be able to complete a complex task alone. One of the best example to customers I use is a banks safe door in old films, with two people with different keys needed to open it

Luca.

[perrenre]

We backup all VM's to local hard drives; as everyone else does. We copy those backups to tape to store offsite (2 week rotation). In addition, we replicate critical VM's to an out of state office via MPLS connection. For business continuity (quick file or VM recovery) we can restore from local disk/tape. For disaster recovery we make some DNS changes and fail over to the replica. This seems to work well for us.

[ddameron]

I voted for "By copying backup files to external hard drive, and taking hard drives off-site", but technically all of my backups are off-site already. Our servers are at a data center, but our VEEAM machine is in our offices. We backup the servers over our PTP connection. However, we do also have rotating USB hard drives that we take home with us as an extra level of protection.

[mschlott]

I off site by having my target in a different data center, but we also back up that data to tape and send it off site. If my target storage fails, I can't aford to not have any backups.

[averylarry]

At most clients I simply have a replica performed every night of every VM to an offsite location. No backup files are ever offsite.

At 1 (paranoid) client I run a replication offsite every night. I also run a replication locally do a 2nd SAN every night. And finally I run a local replication job to an encrypted removable hard drive once every week. I have 5 of those hard drives and they rotate offsite physically.

My clients do not consider a backup history to be a required part of DR or BC. That is why an offsite replication is sufficient.