Safely wipe caching drive to test AWS S3 backup restore

2

2000gtacoma

New Member
Aug 31, 2025
14
2
3
Hi All,

So we are using PBS 4.0.18. Our current setup is our PVE nodes send backups to our proxmox backup server. From there I have a sync job pulling the backups from the datastore on the backup server to our S3 bucket. This seems to be working fine. However, I need to prove that backups from AWS can be directly pulled. The issue I am running into is the backups are also being stored on the caching drive in the backup server. I would like to wipe the caching drive and then do a restore to verifiy the data is being pulled directly from AWS. I understand cost will be involved for pulling the data. So far I can't seem to find anything concrete on this.
 
Hi,
you could either do an s3-refresh followed by a verification (the s3-refresh fetches all metadata, the verify job the chunk contents unconditionally) or re-create the data-store from scratch (without deleting the s3 bucket contents) and use the reuse datastore and overwrite in-use marker flags.
 
Got the following when attempting the refresh

2025-11-18T10:05:05-05:00: Fetching object vm/133/2025-10-15T18:43:57Z/drive-scsi0.img.fidx
2025-11-18T10:05:05-05:00: <?xml version="1.0" encoding="UTF-8"?>
<Error><Code>InvalidObjectState</Code><Message>The operation is not valid for the object's storage class</Message><StorageClass>GLACIER</StorageClass><RequestId>FDZ6FGHBJ9F3ER35</RequestId><HostId>iCX3MWgkX2paOz6HzQP+dUyaa2SnfFXQPUKRwmz0bTUABIUENPrG/lmYUOcNddjl5Zds6q5ti4M=</HostId></Error>
2025-11-18T10:05:05-05:00: TASK ERROR: failed to fetch object vm/133/2025-10-15T18:43:57Z/drive-scsi0.img.fidx: unexpected status code 403 Forbidden
 
So after trying to refresh the s3, the datastore went into maintenance mode. I was able to remove the maintenance mode. But I get the following error during the refresh.

Bad request 400 datastore "my s3 datastore" is unavailable: S3 refresh maintenance mode.

Edit: I deleted the vm in the glacier storage. All seems to be working fine now.
 
Last edited:
2000gtacoma said:
Is my understanding Proxmox is looking at implementing something for glacier or stepped storage?
Click to expand...
Not in the short term

2000gtacoma said:
I was able to remove the maintenance mode. But I get the following error during the refresh.

Bad request 400 datastore "my s3 datastore" is unavailable: S3 refresh maintenance mode.
Click to expand...
This is expected and deliberate. during the refresh no other operations can be performed on the datastore, therefore the maintenance mode is set. If the refresh fails for some reason, user interaction is enforced to check, therefore the maintenance mode is kept as is.
 
Chris said:
Not in the short term


This is expected and deliberate. during the refresh no other operations can be performed on the datastore, therefore the maintenance mode is set. If the refresh fails for some reason, user interaction is enforced to check, therefore the maintenance mode is kept as is.
Click to expand...
Thanks for the quick replies! I'm up and running.
 
You must log in or register to reply here.
Top Bottom