Import existing backups from a new location into HCI

Davide Mauri Lv1Posted Jun-24-2025 22:57

Hello everyone,
    after performing some backup copies to S3 via SCP, we switched to another solution for uploading data to AWS due to the high costs of download traffic performed during the backup windows (*).

Now, if I simulate a disaster scenario and I connect a newly created NFS datastore to HCI (so to mimic the download of backups from S3 on a new device), it seems that these backups are not visibile to HCI. I tried to "Update" the backup repositories, before and after the "Scan for new disks" in the datastore section, but the backups count stays at 0.

Is it something that requires a support request?
I searched in the user guide but I did not found anything like what I'm trying to accomplish. The only somewhat similar post I found is https://community.sangfor.com/forum.php?mod=viewthread&tid=5530 which talks about a session with tech support.

Thanks and best regards.

(*) We had a long session with tech support regarding the high resource usage during S3 metadata checks, and we obtained the S3 patches in advance, but the download usage from AWS remained extremely high.

By solving this question, you may help 792 user(s).

Posting a reply earns you 2 coins. An accepted reply earns you 20 coins and another 10 coins for replying within 10 minutes. (Expired) What is Coin?

Enter your mobile phone number and company name for better service. Go

Davide Mauri Lv1Posted Jul-18-2025 22:16
  
Sorry for coming back on this post so early, but I cannot edit my previous one.
I deleted a clone I created for test, and not the actual VM I used to have inside the backup plans.

So, everything stays correct, apart from seeing the backups of the deleted VM under Reliability > Scheduled Backups/CDP > Backups.
I can see them by browsing the datastore in Reliability > Scheduled Backups/CDP > Backup Repositories.

Sorry for the wrong recap.
Davide Mauri Lv1Posted Jul-18-2025 22:00
  
Hello everyone, I waited a bit to see if there were some changes, but I can confirm that the situation stayed the same.
Today I deleted the source test VM, and I received the usual notification that also those backups would be deleted within 30 days, and I am now seeing this:
- in the recycle bin, residual files section, I actually find the VM backups, with deletion date June 26th (which is the date I removed the source VM from the backup policy)
- in the reliability > backups, I find the "cloned" backups (which I fed HCI via a new NFS repository, detached from backup policies). The latest backup shown says June 5th

So, summing things up, it seems that (but I have to wait a bit more for the expiry timer to kick off):
- if we recover a backup datastore and we present it to HCI, non-overlapping backups (not just the single restore points, but also those which would fall into an existing backup policy) are correctly showing up
- these backups are not "managed" by HCI data retention policies, so if we present the backups of deleted VMs, we can restore those VMs even after 30 days from the deletion
- of course these backups are not managed by backup policies, if we present them on a different datastore

I still cannot speak about what happens if we present VM backups to a newly created HCI cluster, since I cannot simulate a full-blown disaster recovery scenario, but it would seem plausible that we should be able to see these backups and recover the VMs inside them.

It would be nice if a Sangfor representative would give an insight on this.

Anyone else playing around like me on these test?
Apriyanto Lv5Posted Jun-29-2025 21:21
  
If you can only get help through technical support, it would be better if you just follow the instructions.

There may be some things that need to be fixed because there has never been a case like that before.

If the case is successful, then you have helped this community to add knowledge resources if there is a case similar to this

Thank you for sharing
Davide Mauri Lv1Posted Jun-26-2025 18:23
  
Ok, I played around a bit and I noticed this:
- for a recently deleted VM, for which I had backups in the "restored" NFS datastore but *not* in the current backup datastores, I am able to see the restore points
- for an actually running VM, which has a short backup retention and for which I have old backups on the "restored" NFS datastore, I am able to see the "out of retention" restore points
- for the other VMs, for which I have a longer retention (and we have yet to reach the limit), I am unable to see the restore points from the "restored" NFS datastore.

So, to sum up:
- it seems that HCI is actually able to see "imported" backups, retrieved from an external source and put inside a NFS datastore
- is not reindexing them (if that's the correct term), so they are actually present, but not visibile/usable
- so, my assumption is that those backups id somehow collide with existing backups in other datastores (I did not remove the datastore from which I copied those backups from), and datastores used by backup policies correctly take precedence over backups coming from "unused" datastores which may be present, unless the restore points fall outside the retention and so are no more present on the datastores used by backup policies.

I'll talk with the customer to see if I can temporarily remove the "source" backup datastore, so to have proof that what I have just written is correct (as I hope).

I'll let you guys know what I find.
Davide Mauri Lv1Posted Jun-26-2025 15:56
  
Hello AR,
    thanks for the detailed reply. It makes perfect sense, yet (since you mentioned it) Veeam is able to do a "repository rescan" and import/generate backup metadata, assuming that backups are not corrupted.
I was hoping to achieve the same without the intervention of Sangfor technical support, just to be able to perform such disaster recovery tests whenever I wanted to.

I'll play around a bit more with the test NFS share, and see if I missed something, but if everything is correct and I have to interact with tech support, it will be good enough.

I'll just have to let the customer know of this "limitation", but in a real disaster recovery scenario like losing the hypervisors and the local backup storage, involving tech support will be the smallest of the issues to overcome...

Thanks again and best regards.

Davide
AR Lv2Posted Jun-26-2025 01:43
  
Hello,
Thank you for the detailed context — this is a fairly advanced scenario involving disaster recovery in a Sangfor HCI environment.

Summary of Issue:
You're trying to manually re-attach an NFS backup repository (previously copied from S3) to Sangfor HCI, expecting it to detect and index existing backup files, but it's showing 0 backups, even after:

Updating the backup repository

Scanning for new disks

Key Point:
Sangfor HCI does not automatically re-index or recognize backups copied from an object storage like S3 (or a backup-targeted NFS) unless those backups are registered through its internal backup metadata and database. The backup system in Sangfor HCI/VMP is tightly integrated with its internal metadata management, which tracks:

Backup chain integrity

Indexing

Backup repository mapping

Why You Can't See the Backups:
When you upload/download backup files outside the Sangfor backup engine workflow (e.g. manually via SCP or S3 tools), HCI:

Has no metadata mapping for those files

Won't list them in the repository even after disk scans

This is by design, to ensure backup consistency and integrity — similar to how Veeam or other backup systems require catalog files to access backup chains.

Recommended Next Steps:
✅ Option 1: Engage Sangfor Support (Best)
Yes, your assumption is correct — this typically requires intervention from Sangfor technical support, especially if:

You want to import and use these backups on a clean/reinstalled HCI system

The metadata database is missing or not restored

As referenced in the forum link you mentioned (thread 5530), they can provide tools or manual commands to rebuild or import backup metadata from the NFS files — if they are still usable and intact.

✅ Option 2: Restore from Backups Manually (if urgent)
If you're in a disaster recovery scenario and just need the data urgently, and can’t wait for support:

You may be able to manually mount and extract the VM disks from the backup files on the NFS share (assuming you know the format — e.g. .vhd, .qcow, or Sangfor proprietary)

But this is not officially supported and may require deeper understanding of the Sangfor backup structure.

Doll Lv1Posted Jun-25-2025 03:43
  
The issue you're experiencing with Sangfor HCI not recognizing backups on a newly connected NFS datastore might be due to several reasons:

Potential Causes
1. *NFS Configuration*: Ensure that the NFS datastore is correctly configured and mounted on the HCI nodes. Verify that the NFS server is exporting the correct directory and that the HCI nodes have the necessary permissions to access it.
2. *Backup Repository Configuration*: Check that the backup repository is correctly configured in Sangfor HCI. Ensure that the repository type is set to NFS and that the correct NFS datastore is selected.
3. *Scan and Update*: After connecting the new NFS datastore, perform a "Scan for new disks" in the datastore section, followed by an "Update" on the backup repositories. This should refresh the list of available backups.

Troubleshooting Steps
1. *Verify NFS Connectivity*: Verify that the HCI nodes can connect to the NFS server and access the exported directory.
2. *Check Backup Repository Logs*: Check the backup repository logs for any errors or warnings that might indicate why the backups are not being recognized.
3. *Contact Support*: If you've tried the above steps and the issue persists, consider reaching out to Sangfor's technical support for further assistance.

Additional Considerations
1. *Backup Format*: Ensure that the backups are in a format that Sangfor HCI can recognize and restore from.
2. *NFS Version*: Verify that the NFS version used by the HCI nodes is compatible with the NFS server.

By investigating these potential causes and performing the troubleshooting steps, you should be able to resolve the issue with Sangfor HCI not recognizing backups on the newly connected NFS datastore.

I Can Help:

Change

Board Leaders

NyxZale...

Weekly Sharers

Newbie5...

Weekly Questioners