spped comparison NDMP to Tape vs. NAS to S3

[HansMeiser]

Hello,

currently i do some tests with larger netapp-volumes.
backup 1 ist netapp volume in size of 20T backed up by NDMP to tape. we need ~18h to run backup successfully. bottleneck is tapelibrary.

backup 2 is same volume using NAS-Backup to S3. bottleneck is source/netapp

To my surprise currently we have already 18 hours reached and still 25% to go. i expected storing to s3 would be much faster.

One factor is our av-scanner. it seems running an NDMP backup not uses av-scanner. by switching to NAS backup now our backupserver sees a lot of files and every single file is scanned. so we have a significant time overhead in comparisation between ndmp+tape vs. nas+s3.
What would happen if we disable av-svanner and nas to s3 can run without encroachments? would it be as fast as ndmp to tape or are there other limitations?

What is your experience in storing large shares to different targets?

Thanks,
Hans

[Mildur]

Hi Hans,

You’re comparing different technologies — how we retrieve data (NDMP vs. file backup engine using SMB or NFS) and the backup target (tape vs. S3 object storage).
I can’t say whether you’ll achieve the same performance as NDMP when disabling antivirus scans, though I don’t expect it. However, it will perform faster without antivirus scans. And our file backup engine will give you many benefits over NDMP, such as Granular restore or Instant Recover capabilities.

I recommend running a test with antivirus disabled on your file proxy.
If you find the backup performance still too slow, please contact our customer support to investigate potential bottlenecks.

Best regards,
Fabian

[HansMeiser]

Hello Fabian,

yes, we compare two different technologies. but time is in a comprehensive view also an important factor.
AS i read full-backups are only triggered if backup chain is broken. so in an ideal case we have one fullbackup at the beginning and lifetime inkrementell backups.

I set some folders and unc paths as exclusions for ms-defender. it seems it is not scanning when starting a new inkrementell file-share backup.

How can i trigger a new full backup? Is is enough to delete the backups under "object storage"?

Thanks,
Hans

[Mildur]

Hi Hans

Yes, that would trigger a new Active Full on the next job run.
Or create a new Job (you can clone the existing one).

Best,
Fabian