[pbs-devel] [RFC v2 proxmox-backup 02/21] datastore: mark groups as trash on destroy

Fabian Grünbichler f.gruenbichler at proxmox.com
Fri May 9 14:27:51 CEST 2025


On May 8, 2025 3:05 pm, Christian Ebner wrote:
> In order to implement the trash can functionality, mark all the
> snapshots of the group and the group itself as trash instead of
> deleting them right away. Cleanup of the group is deferred to the
> garbage collection.
> 
> Groups and snapshots are marked by the trash marker file. New backups
> to this group will check for the marker file (see subsequent
> commits), clearing the whole group and all of the snapshots to
> create a new snapshot within that group. Otherwise ownership
> conflicts could arise. This implies that a new backup clears the
> whole trashed group.

this seems a bit surprising.. couldn't we check the new and older owner,
and abort if there's a mismatch, but proceed otherwise?

the implementation of this also doesn't happen in this patch though, so
maybe drop it here in any case - this just implements trashing groups..

> Snapshots already marked as trash within the same backup group will
> be cleared as well when the group is requested to be destroyed with
> skip trash.

this makes sense

> 
> Signed-off-by: Christian Ebner <c.ebner at proxmox.com>
> ---
>  pbs-datastore/src/backup_info.rs | 19 ++++++++++++++++---
>  pbs-datastore/src/datastore.rs   |  4 ++--
>  2 files changed, 18 insertions(+), 5 deletions(-)
> 
> diff --git a/pbs-datastore/src/backup_info.rs b/pbs-datastore/src/backup_info.rs
> index 76bcd15f5..9ce4cb0f8 100644
> --- a/pbs-datastore/src/backup_info.rs
> +++ b/pbs-datastore/src/backup_info.rs
> @@ -215,7 +215,7 @@ impl BackupGroup {
>      ///
>      /// Returns `BackupGroupDeleteStats`, containing the number of deleted snapshots
>      /// and number of protected snaphsots, which therefore were not removed.
> -    pub fn destroy(&self) -> Result<BackupGroupDeleteStats, Error> {
> +    pub fn destroy(&self, skip_trash: bool) -> Result<BackupGroupDeleteStats, Error> {
>          let _guard = self
>              .lock()
>              .with_context(|| format!("while destroying group '{self:?}'"))?;
> @@ -229,14 +229,20 @@ impl BackupGroup {
>                  delete_stats.increment_protected_snapshots();
>                  continue;
>              }
> -            snap.destroy(false, false)?;
> +            snap.destroy(false, skip_trash)?;
>              delete_stats.increment_removed_snapshots();
>          }
>  
>          // Note: make sure the old locking mechanism isn't used as `remove_dir_all` is not safe in
>          // that case
>          if delete_stats.all_removed() && !*OLD_LOCKING {
> -            self.remove_group_dir()?;
> +            if skip_trash {
> +                self.remove_group_dir()?;
> +            } else {
> +                let path = self.full_group_path().join(TRASH_MARKER_FILENAME);
> +                let _trash_file =
> +                    std::fs::File::create(path).context("failed to set trash file")?;
> +            }
>              delete_stats.increment_removed_groups();
>          }
>  
> @@ -245,6 +251,13 @@ impl BackupGroup {
>  
>      /// Helper function, assumes that no more snapshots are present in the group.
>      fn remove_group_dir(&self) -> Result<(), Error> {
> +        let trash_path = self.full_group_path().join(TRASH_MARKER_FILENAME);
> +        if let Err(err) = std::fs::remove_file(&trash_path) {
> +            if err.kind() != std::io::ErrorKind::NotFound {
> +                bail!("removing the trash file '{trash_path:?}' failed - {err}")
> +            }
> +        }
> +
>          let owner_path = self.store.owner_path(&self.ns, &self.group);
>  
>          std::fs::remove_file(&owner_path).map_err(|err| {
> diff --git a/pbs-datastore/src/datastore.rs b/pbs-datastore/src/datastore.rs
> index 6df26e825..e546bc532 100644
> --- a/pbs-datastore/src/datastore.rs
> +++ b/pbs-datastore/src/datastore.rs
> @@ -581,7 +581,7 @@ impl DataStore {
>          let mut stats = BackupGroupDeleteStats::default();
>  
>          for group in self.iter_backup_groups(ns.to_owned())? {
> -            let delete_stats = group?.destroy()?;
> +            let delete_stats = group?.destroy(true)?;
>              stats.add(&delete_stats);
>              removed_all_groups = removed_all_groups && delete_stats.all_removed();
>          }
> @@ -674,7 +674,7 @@ impl DataStore {
>      ) -> Result<BackupGroupDeleteStats, Error> {
>          let backup_group = self.backup_group(ns.clone(), backup_group.clone());
>  
> -        backup_group.destroy()
> +        backup_group.destroy(true)
>      }
>  
>      /// Remove a backup directory including all content
> -- 
> 2.39.5
> 
> 
> 
> _______________________________________________
> pbs-devel mailing list
> pbs-devel at lists.proxmox.com
> https://lists.proxmox.com/cgi-bin/mailman/listinfo/pbs-devel
> 
> 
> 




More information about the pbs-devel mailing list