Question(s) about veeam usability for different scenarios

[Daneels]

Hi All, I am new to the board and have several questions (I hope this post is not too long and confusing, but I am just looking for extra info to (hopefully) make a good judgement about using veeam or not)).

Our VM setup

• 4 ESXi hosts
• SAN storage shared between hosts for VM's in production
• 47 VM's (taking up about 3Tb) so 75Gb on 'average'
• However real distribution is more like
  • 7 machines <15Gb
  • 18 machines <45Gb
  • 11 machines <100Gb
  • 9 machines <200Gb
  • 1 machine 360 Gb
  • 1 machine 2.8 Tb but all of those volumes (largest is 1Tb) (except OS) are RDM (and in addition there is 20Tb archive volumes via iSCSI)

Current setup for backup (VM and file) / replication…

vRanger

• We are making replica's of our VM's to an older SAN based storage for all of the VM's (except the 2.8Tb machine) on a daily basis. 2 database servers are replicated on a 6 hour scheme.
• In addition to this we make backups of our VM's to 2 NAS storage devices on a weekly basis (first week of month full backup, rest of the month incremental)
• These backups of the VM's are later written to tape (using Netvault backup software (see further)

Netvault

• In addition to vRanger we are using Netvault for backing up
  • The RDM volumes (weekly full, daily incremental) 6 weeks retention
  • The volumes of the second largest file server (weekly full, daily incremental) 6 weeks retention
  • Backup up the 'system state' of a number of vm's
  • These backups are done to VTL (only last week is kept there) and duplicated onto tape from that location.
• It is also used to create TAPE archives from the 20Tb of archive volumes (two copies actually).
• When these archive volumes are 'full' a final third copy is made (and it is kept in the tape library.

So for so good, however:

• Both vRanger and Netvault have gone through a change in ownership: 2 separate companies -> Quest software -> DELL
• And the sales / support channels in this part of the world has been less than spectacular with all these changes.
• Our infrastructure is showing its age (e.g tape library is close to 6 years)
• A number of new possibilities are cropping up (e.g. Cloud storage, VTL alike appliances…)

And now for the questions:

• Need for tape? Do any of you have a complete tapeless setup? If so how do you implement the 3-2-1 rule (3 copies, 2 media types, 1 off site)?
• Would it be a feasible scenario to cover my backup / replication needs with veeam only (replacing both vRanger and Netvault)? With or without tape?
• Do any of you have experience with BIG machines (like the 2.8Tb one) and would you even consider to replicate this machine? Or make of backup of the machine (instead of a file system backup of the volumes like we do today)
• And finally (bit OT)
  • If I take away the Netvault product entirely I am still stuck with the archive data.
  • Would you consider using veeam for these type of file backup?
  • The 20Tb volume is typically sliced up in yearly datasets of 4-5Tb?
  • Anyone out there that would be using the cloud for example as a secondary / third location for this data
  • A back of the envelope calculation gave me a price of 10500 euros to store 20Tb of data for 5 years in Amazon Glacier, while buying a 24 slot LTO6 library with 2 drives and 5 years of support and the necessary cartridges for dual copies would set me back 15000 euro?

So any thoughts on this?

(perhaps a bit much for one post but it is all linked hence one post)

[foggy]

Peter, welcome to the Veeam Community Forums!

• Need for tape? Do any of you have a complete tapeless setup? If so how do you implement the 3-2-1 rule (3 copies, 2 media types, 1 off site)?

I believe there is a large number of tapeless setups among our customers, since the 3-2-1 rule can be basically met, for example, by having local backups, remote replicas, and offsite backups (cloud or simply rotated drives).

[*]Would it be a feasible scenario to cover my backup / replication needs with veeam only (replacing both vRanger and Netvault)? With or without tape?

I would say yes, probably except the archive volumes, since they are iSCSI. What kind of RDM volumes do you use?

[*]Do any of you have experience with BIG machines (like the 2.8Tb one) and would you even consider to replicate this machine? Or make of backup of the machine (instead of a file system backup of the volumes like we do today)

If you search these forums, you will find a lot of existing threads discussing huge VMs backup (even much larger than yours) with no issues. Here's one of them: What size is the largest vm you backup ?

Hope this helps.

[Daneels]

Hi, thx for the quick reply.

I would say yes, probably except the archive volumes, since they are iSCSI. What kind of RDM volumes do you use?

The big server has got

• 900Gb RDM (which sits on SAN storage)
• 800Gb RDM (also on SAN)
• 100Gb RDM (on SAN)
• 500Gb RDM (on SAN)

they are all physical compatibility mode.

And next to those active work volumes there are 8 archive volumes mounted via iSCSI (each 2.5Tb in size)
These volumes sit on cheap NAS storage and are presented via iSCSI.

Hope this clarifies a bit.

Is it correct that you can selectively backup specific vmdk's via veeam?

Peter

[foggy]

Is it correct that you can selectively backup specific vmdk's via veeam?

Yes, that's correct.

However, neither physical RDM nor volumes presented via iSCSI are supported by Veeam B&R (basically, you cannot snapshot them, so they are simply skipped during backup).

[Daneels]

So I would need to make the volumes by creating vmdk's on vmfs volumes? That would allow me to backup those.

The archive volumes (iSCSI) wouldn't be backed up anyhow (well at least not via this method)

[veremin]

Yes, in order for VB&R to backup given volumes, the volumes should be of supported type. Meanwhile, those connected via ISCSI initiator will be skipped from processing. Thanks.