Yes, looks like an error returned from backblaze. But in that case stopping the task and starting again shouldn't clear that?
It could, as the connection is reset and a new connection established in between backblaze API and PBS s3 client.Yes, looks like an error returned from backblaze. But in that case stopping the task and starting again shouldn't clear that?
2025-08-14T06:13:04+00:00: verify backblaze:ct/103/2025-04-20T08:28:37Z
2025-08-14T06:13:04+00:00: check pct.conf.blob
2025-08-14T06:13:04+00:00: check root.ppxar.didx
2025-08-14T06:16:48+00:00: "can't verify chunk, load failed - client error (SendRequest)"
2025-08-14T06:16:50+00:00: corrupted chunk renamed to "/mnt/cache/.chunks/9b04/9b0470ce8a908c5695dc83c70f9c13f0aa5d007db5a676b43cf0ea9ef72b9b39.0.bad"
this s3 client error should indeed not lead to the locally cached chunk being renamed. This has also been reported in https://bugzilla.proxmox.com/show_bug.cgi?id=6665, please subscribe there to get status updates on this issue.Is that a network error? Or is the chunk actually bad?Code:2025-08-14T06:13:04+00:00: verify backblaze:ct/103/2025-04-20T08:28:37Z 2025-08-14T06:13:04+00:00: check pct.conf.blob 2025-08-14T06:13:04+00:00: check root.ppxar.didx 2025-08-14T06:16:48+00:00: "can't verify chunk, load failed - client error (SendRequest)" 2025-08-14T06:16:50+00:00: corrupted chunk renamed to "/mnt/cache/.chunks/9b04/9b0470ce8a908c5695dc83c70f9c13f0aa5d007db5a676b43cf0ea9ef72b9b39.0.bad"
Maybe I should change my backblaze account to EU instead of US (I am based in EU and made that account before EU was an option).
Hm, did I miss something but do I understand correctly that client-encrypted backups (like described on https://pve.proxmox.com/wiki/Storage:_Proxmox_Backup_Server#storage_pbs_encryption ) are encrypted in any case, even if the first datastore used by the PBS is on a S3 storage? I mean the feature request is about adding server-level encryption (e.g. for encrypting a datastore before offsite replication) while the wiki-page is about client-level encryption.No, this is currently not planned, but given that datastore plain text datastore contents can be synced to a potentially less or un-trusted provider, it could make sense to add an additional server side encryption layer. Note however that there are also server side encryption options for buckets. But do open an enhancement request for this at https://bugzilla.proxmox.com linking this post (the thread itself is to crowded to be liked as a whole).
Yes, client side encrypted snapshots are and will be encrypted also on the s3 backend. As you understood correctly, the request is to add server side encyrption (performed by the PBS host before sending it to S3). Currently unencrypted snapshots are stored in plain text in the s3 bucket, only the transfer is encrypted by using TLS.Hm, did I miss something but do I understand correctly that client-encrypted backups (like described on https://pve.proxmox.com/wiki/Storage:_Proxmox_Backup_Server#storage_pbs_encryption ) are encrypted in any case, even if the first datastore used by the PBS is on a S3 storage? I mean the feature request is about adding server-level encryption (e.g. for encrypting a datastore before offsite replication) while the wiki-page is about client-level encryption.
You need to set theHi, I try to mount a S3 bucket from Backblaze, but still have:
access time safety check failed: failed to upload chunk to s3 backend: chunk upload failed: unexpected status code 501 Not Implemented
Proxmox Backup Server is the latest stable, with no subscription
Skip If-None-Match header
in the provider quirks (advanced options for the endpoint create/edit window). Unfortunately not all providers implement the conditional upload.We use essential cookies to make this site work, and optional cookies to enhance your experience.