Seems obvious, but want confirmation

[SecondCity]

Hi all,

Extremely obvious-sounding question here, but I don't recall seeing any confirmation in the User Guide (or on this forum...yet):

Is Veeam running a backup on a snapped VM? Meaning, of course, that data can continue to change on the server and Veeam won't be capturing those change (or alternatively causing conflicts)? I only ask because I have a very large server that took 36 hours to complete its first backup the other day. I'm taking steps to speed up the process, but fear I'll still have a pretty long wait on a full backup.

Or is my assumption incorrect, and Veeam is running on the production server, leaving the snapshot as just a rollback point in case something were to go wrong?

Thanks very much!
John

[TrevorBell]

John,

First backup may take that long my 921gig file print server took about 30 hours on the first pass, but with Vsphere v4 upwards and CNT change block tracking all subsequent backups will be quicker as it will only backup changed data,

[SecondCity]

Yes, I've seen that, for sure (in fact, I'm running into some resource problems for incrementals...looking at the Microsoft KB article posted in a different thread on the forum here)! Still curious as to how Veeam does it's job, though: does it back up the snapped VM or does it just snap in order to restore in case of error?

[TrevorBell]

backup the snapshot

[SecondCity]

Right on, thanks. So other than causing a little extra traffic on the iSCSI bus, running a backup midday isn't a big deal, huh?

[Neurobit]

Right on, thanks. So other than causing a little extra traffic on the iSCSI bus, running a backup midday isn't a big deal, huh?

Second,
I run backups during the day with no perceived drop in performance. Three concurrent jobs twice a day (at 10 AM & 7:00 PM).
We have a FC SAN though. This is where we have our ESX cluster storage. Backup target is Gig Ethernet NAS.

[SecondCity]

Hey Neurobit,

Great to hear! Would you mind elaborating on which Gig Ethernet NAS you use? We're trying to evaluate where to target our backups. Also, is your Veeam server virtualized, or is it physical?

Thanks!
John

[Neurobit]

Hey Neurobit,

Great to hear! Would you mind elaborating on which Gig Ethernet NAS you use? We're trying to evaluate where to target our backups. Also, is your Veeam server virtualized, or is it physical?

Thanks!
John

John,

Our Veeam server is running on a virtual 2008 R2.

We have a Unistar NAS. It's about 5 years old, so I am actually looking at getting a faster one, since it seems to lag even though I have both Gig interfaces bonded. To give you an idea how painful it is at times, I did my first backup to tape this weekend (copied all my VRB/VRK from the NAS to tape), and it took 26 hours for ~1.5 TB using Backup Exec (1 GB/Min, or around 17 MB/S). I have seen our tape backups go a lot faster than this, so the bottleneck seems to be the NAS.

I am new to Veeam (I've deployed it about two weeks ago), and so far I am pretty impressed. For physical servers I am running Backup Exec 12.5.

I am still playing with jobs and multiple VMs, and even though you get better dedup the more VMs you backup, I'd rather run concurrent jobs to shrink my backup window.

Right now I have a job that contains a DC and our big file server (1.5 TB), a job for Exchange, and a job for our BES server. They all run concurrently without issues twice a day. My big file server averages a processing rate of between 200-400 MB/s, so it's usually backed up in about 1 to 1.5 hours. Our BES server takes 3 minutes (70 GB @ 250MB/S), and Exchange 3 to 3.5 hours since the processing rate is so slow (21 MB/s today for a 248GB VM).

Thanks - Paul

[TrevorBell]

I run jobs anytime of the day if im testing with no performance drop at all, backing up around 40 vm`s