Best practice in configuring Veeam Backup repository in DAS

[albertwt]

People,

I have just installed the new 36 TB Disk enclosure to my existing Backup Repository host running VBR 9.0 Update 1 on Win 2008 R2 OS.

So I wonder what is the best practice in configuring and setting the new LUN to accommodate the best practice in configuring it according to Veeam ?
eg. LUN size, stripe size, NTFS Allocation Unit Size, etc...

At the moment there are another drives that is 10-12 TB LUNs already working as NTFS drives.

Thanks,

[albertwt]

My server and the storage enclosure is HP brand.

I wonder if there is updated article http://veeampdf.s3.amazonaws.com/whitep ... rmance.pdf for Scale Out Backup Repo or Veeam 9.0 U1 ?

[PTide]

Hi,

LUN size

Depends on how long your backup is.

stripe size, NTFS Allocation Unit Size

Please check this thread.

You might want to check A Deep Look at Scale-out Backup Repository.

Thanks

[albertwt]

Hi,

Please check this thread.

You might want to check A Deep Look at Scale-out Backup Repository.

Thanks

Cool,

Many thanks Mr. Tide, I will select the largest Strip size possible supported by my Storage Controller.

[mcrape]

I highly recommend watching the 2015 Backup Repository Best Practices video by Gostev. A lot of good info in there, including tidbits about block sizes, allocation units, and RAID penalties.

If you end up using dedupe, Server 2016 is apparently worth the wait as dedupe is much better in it vs. 2012 R2.

[ferrus]

During our Veeam server installs, I amended the NTFS alloc size of our 43TB DAS repos according to best practice (can't remember what it is off hand).

Months later when we started Instant Restore testing, we found that the DAS storage couldn't be used for the vPower NFS folder.
The c: drive and SSD worked perfectly fine, however.
It was suggested by Veeam support that there may be an incompatibility between VMware's NFS implementation and NTFS block size.

[albertwt]

During our Veeam server installs, I amended the NTFS alloc size of our 43TB DAS repos according to best practice (can't remember what it is off hand).

Months later when we started Instant Restore testing, we found that the DAS storage couldn't be used for the vPower NFS folder.
The c: drive and SSD worked perfectly fine, however.
It was suggested by Veeam support that there may be an incompatibility between VMware's NFS implementation and NTFS block size.

So what is the suggested maximum Allocation unit size ?

Can you please share it here with the below command:

Code: Select all

fsutil fsinfo ntfsinfo D:

or any other drive that is used for vPower NFS service ?
fsutil fsinfo ntfsinfo drive letter:

[ferrus]

Think the 43TB DAS was setup with 64k, and the 980GB SSD with the standard 4k.
The DAS was formatted with the large file switch, as advised.

Code: Select all

RAID 6 DAS (43TB)

NTFS Version   :                  3.1
LFS Version    :                  2.0
Number Sectors :                  0x00000002ba7cddff
Total Clusters :                  0x000000002ba7cddf
Free Clusters  :                  0x0000000009b0d69b
Total Reserved :                  0x0000000000000000
Bytes Per Sector  :               4096
Bytes Per Physical Sector :       4096
Bytes Per Cluster :               65536
Bytes Per FileRecord Segment    : 4096
Clusters Per FileRecord Segment : 0


SSD

NTFS Version   :                  3.1
LFS Version    :                  2.0
Number Sectors :                  0x000000006fbd0fff
Total Clusters :                  0x000000000df7a1ff
Free Clusters  :                  0x000000000dbf870a
Total Reserved :                  0x0000000000000000
Bytes Per Sector  :               512
Bytes Per Physical Sector :       512
Bytes Per Cluster :               4096
Bytes Per FileRecord Segment    : 1024
Clusters Per FileRecord Segment : 0 

[albertwt]

Cool, many thanks for the information and the sharing Ferrus.