Is my Veeam Backup slow or fast?

[baatch]

Hi All

I'm a new Veeam user and just setup my first Veeam backup. Just wanted to know if my Backup is going slow or fast. If anyone can help comment on the stats below:


Active Full Backup
Success 56
Start time 22:00:00
Total size 8,6 TB
Backup size 3,1 TB
Warning 0
End time 08:29:03 (+1)
Data read 3,7 TB
Dedupe 1,1x
Error 0
Duration 10:29:02
Transferred 3,3 TB
Compression 1,1x

Processing rate: 106 MB/S
Bottleneck: Source

-----------------------------------------------------------------------------------------------------------
Incremental
Success 56
Start time 22:00:11
Total size 8,6 TB
Backup size 157,0 GB
Warning 0
End time 00:11:26 (+1)
Data read 186,1 GB
Dedupe 1,1x
Error 0
Duration 2:11:14
Transferred 158,0 GB
Compression 1,1x

Processing rate: 999 MB/s
Bottleneck: Source

[foggy]

Processing rate: 106 MB/S
Bottleneck: Source

Pretty good pace, I would say. Alexander, could you please tell us a bit about your setup so we could review it and probably give some recommendations?

[baatch]

Backup server is a HP DL180 G6, 12 GB RAM, 7x3TB SATA, HP FC HBA 4GB, Windows Server 2012 with Dedupe, Veeam 6.5 with Direct SAN access.

Vmware Vsphere 5.1, HP 3PAR Storeserv 7200 connected to 8GB FC switches.

Are the backup speeds normal for this setup or should I be able to go faster?

[chrisdearden]

I've seen a little faster on the full backup , but its certainly acceptable. How many machines are in the Job?

[baatch]

56 VM's, is there anything more I could configure to get it faster?

[chrisdearden]

It looks like you are processing a VM every 2-3 minutes , which is pretty good. Have you just got the single job for your whole environment ?

[baatch]

Yes only 1 single job for all of the VM's. Should I divide them into 2 jobs instead?

[chrisdearden]

That would probably be a better plan. 3-4TB of Source VM's per job is a nice point to aim for in the current version. If you are going to attempt 2012 dedupe , you might have a bit of a challenge processing that much data on a single volume.

What sort of retention are you aiming for ?

[baatch]

I don't know if my strategy is wrong but I'm thinking that Primary backup server should have 30 restore points, forward incremental, active full backup each saturday.
Last full backup of the month will replicate to another server in another building and also to tapes to take offsite.

Any suggestion on my strategy and how I should configure it to 2 backup jobs instead ?

Can't wait till v7 so I can have the built in backup copy job, gfs feature and tape !

[Yuki]

the speed is OK, but not great. Still better than what we are seeing, yet given your storage - i would expect better speeds.

Your incremental runs at about 20MB/s if I'm reading it correctly (158GB over 2:11 h:m).

[foggy]

Any suggestion on my strategy and how I should configure it to 2 backup jobs instead ?

Your strategy looks good. What Chris is talking about is you can split your job into several ones (for example, backup 28 VMs in one job and other 28 in another job with the same retention settings) to be able to run them concurrently. This will increase the overall processing rate.

[Yuki]

But this will also potentially cause the backups to be larger in sum, as deduplication will be done on two smaller sets and those two sets will not share any blocks. So even identical blocks in both jobs will be sent to the backup files.

[dellock6]

Maybe, but deduplication factor cannot be the only goal in setting a job. Backup file size and execution cuncurrency are other important elements. It depends how much importance you give to these...

Luca.

[Vitaliy S.]

Also getting back to the deduplication concern...when you trigger incremental job runs, usually blocks that are transferred as changed ones are new and unique, so I wouldn't expect high deduplication rates for incremental runs.

[baatch]

So the dedup for my full and incremental will decrease.. Hard to decide.

Is there a maximum file size for the Windows Server 2012 dedupe?

[Vitaliy S.]

Not sure there is any maximum for file size if you want to use Window Server 2012 dedupe feature, however there are some limits for the amount of data that Windows Server 2012 can process during the day. Please take a look at this topic for further reading > Best Practice for MS Server 2012 DeDup Repo