From: "Fabian Grünbichler" <f.gruenbichler@proxmox.com>
To: Proxmox Backup Server development discussion
<pbs-devel@lists.proxmox.com>
Subject: [pbs-devel] superseded: [PATCH proxmox-backup] gc: only create locally missing chunk markers if present on S3 store
Date: Tue, 11 Nov 2025 12:46:01 +0100 [thread overview]
Message-ID: <1762861538.w8u0ks6jw1.astroid@yuna.none> (raw)
In-Reply-To: <20250917134131.444585-1-c.ebner@proxmox.com>
a variant of this will be included in the next s3-race-fix patch series
version ;)
On September 17, 2025 3:41 pm, Christian Ebner wrote:
> This fixes a logic issue with the local cache marker files for all
> indexed chunks being created during phase 1 of the garbage collection,
> independent on whether the chunks might have been moved to be bad by
> a verification job.
>
> Therefore, check if the chunk exists in the S3 object store before
> creating the marker, if no bad chunk marker is present in the local
> store cache (this can also happen after reusing a fresh local store
> cache after a new datastore setup).
>
> Make sure to also keep the bad chunk marker files in the local store
> cache. This is achieved by marking these unconditionally if the chunk
> marker file is not found.
>
> Signed-off-by: Christian Ebner <c.ebner@proxmox.com>
> ---
> Encountered while investigating an unrelated issue.
>
> pbs-datastore/src/datastore.rs | 73 +++++++++++++++++++---------------
> 1 file changed, 41 insertions(+), 32 deletions(-)
>
> diff --git a/pbs-datastore/src/datastore.rs b/pbs-datastore/src/datastore.rs
> index 7cf020fc0..4562b0fca 100644
> --- a/pbs-datastore/src/datastore.rs
> +++ b/pbs-datastore/src/datastore.rs
> @@ -1308,40 +1308,49 @@ impl DataStore {
> }
> }
>
> - match s3_client {
> - None => {
> - // Filesystem backend
> - if !self.inner.chunk_store.cond_touch_chunk(digest, false)? {
> - let hex = hex::encode(digest);
> - warn!(
> - "warning: unable to access non-existent chunk {hex}, required by {file_name:?}"
> - );
> -
> - // touch any corresponding .bad files to keep them around, meaning if a chunk is
> - // rewritten correctly they will be removed automatically, as well as if no index
> - // file requires the chunk anymore (won't get to this loop then)
> - for i in 0..=9 {
> - let bad_ext = format!("{}.bad", i);
> - let mut bad_path = PathBuf::new();
> - bad_path.push(self.chunk_path(digest).0);
> - bad_path.set_extension(bad_ext);
> - self.inner.chunk_store.cond_touch_path(&bad_path, false)?;
> - }
> + if !self.inner.chunk_store.cond_touch_chunk(digest, false)? {
> + let hex = hex::encode(digest);
> + let mut is_bad_chunk = false;
> +
> + // touch any corresponding .bad files to keep them around, meaning if a chunk is
> + // rewritten correctly they will be removed automatically, as well as if no index
> + // file requires the chunk anymore (won't get to this loop then)
> + for i in 0..=9 {
> + let bad_ext = format!("{}.bad", i);
> + let mut bad_path = PathBuf::new();
> + bad_path.push(self.chunk_path(digest).0);
> + bad_path.set_extension(bad_ext);
> + if self.inner.chunk_store.cond_touch_path(&bad_path, false)? {
> + is_bad_chunk = true;
> }
> }
> - Some(ref _s3_client) => {
> - // Update atime on local cache marker files.
> - if !self.inner.chunk_store.cond_touch_chunk(digest, false)? {
> - let (chunk_path, _digest) = self.chunk_path(digest);
> - // Insert empty file as marker to tell GC phase2 that this is
> - // a chunk still in-use, so to keep in the S3 object store.
> - std::fs::File::options()
> - .write(true)
> - .create_new(true)
> - .open(&chunk_path)
> - .with_context(|| {
> - format!("failed to create marker for chunk {}", hex::encode(digest))
> - })?;
> +
> + match s3_client {
> + None => warn!(
> + "warning: unable to access non-existent chunk {hex}, required by {file_name:?}"
> + ),
> + Some(ref s3_client) => {
> + if !is_bad_chunk {
> + // Might be a new empty cache with no markers present yet, so fallback to
> + // check the s3 object store in that case.
> + let object_key = crate::s3::object_key_from_digest(digest)?;
> + let head_object_response =
> + proxmox_async::runtime::block_on(s3_client.head_object(object_key))
> + .context("failed to check chunk presence in s3 object store")?;
> + if head_object_response.is_some() {
> + info!("Create missing marker for existing chunk {hex}");
> + let (chunk_path, _digest) = self.chunk_path(digest);
> + // Insert empty file as marker to tell GC phase2 that this is
> + // a chunk still in-use, so to keep in the S3 object store.
> + std::fs::File::options()
> + .write(true)
> + .create_new(true)
> + .open(&chunk_path)
> + .with_context(|| {
> + format!("failed to create marker for chunk {hex}")
> + })?;
> + }
> + }
> }
> }
> }
> --
> 2.47.3
>
>
>
> _______________________________________________
> pbs-devel mailing list
> pbs-devel@lists.proxmox.com
> https://lists.proxmox.com/cgi-bin/mailman/listinfo/pbs-devel
>
>
>
_______________________________________________
pbs-devel mailing list
pbs-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pbs-devel
next prev parent reply other threads:[~2025-11-11 11:45 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-09-17 13:41 [pbs-devel] " Christian Ebner
2025-11-11 11:46 ` Fabian Grünbichler [this message]
2025-11-11 14:31 ` [pbs-devel] obsolete: " Christian Ebner
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1762861538.w8u0ks6jw1.astroid@yuna.none \
--to=f.gruenbichler@proxmox.com \
--cc=pbs-devel@lists.proxmox.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox