From: Christian Ebner <c.ebner@proxmox.com>
To: pbs-devel@lists.proxmox.com
Subject: [pbs-devel] [PATCH proxmox-backup 1/6] verify/datastore: make rename corrupt chunk a datastore helper method
Date: Thu, 16 Oct 2025 15:18:14 +0200 [thread overview]
Message-ID: <20251016131819.349049-2-c.ebner@proxmox.com> (raw)
In-Reply-To: <20251016131819.349049-1-c.ebner@proxmox.com>
By making this a helper of the datastore, within this method it will
become possible to access the inner chunk store for locking ecc.
That will be required to correctly lock the store to avoid
concurrent chunk inserts and garbage collection operations during the
rename, to guarantee consistency on datastores with s3 backend.
Signed-off-by: Christian Ebner <c.ebner@proxmox.com>
---
pbs-datastore/src/datastore.rs | 70 +++++++++++++++++++++++++++++++
src/backup/verify.rs | 75 +---------------------------------
2 files changed, 72 insertions(+), 73 deletions(-)
diff --git a/pbs-datastore/src/datastore.rs b/pbs-datastore/src/datastore.rs
index 038306166..802a39536 100644
--- a/pbs-datastore/src/datastore.rs
+++ b/pbs-datastore/src/datastore.rs
@@ -2418,4 +2418,74 @@ impl DataStore {
.map_err(|err| format_err!("{err:#}"))?;
Ok((backend_type, Some(s3_client)))
}
+
+ pub fn rename_corrupted_chunk(&self, digest: &[u8; 32]) {
+ let (path, digest_str) = self.chunk_path(digest);
+
+ let mut counter = 0;
+ let mut new_path = path.clone();
+ loop {
+ new_path.set_file_name(format!("{}.{}.bad", digest_str, counter));
+ if new_path.exists() && counter < 9 {
+ counter += 1;
+ } else {
+ break;
+ }
+ }
+
+ let backend = match self.backend() {
+ Ok(backend) => backend,
+ Err(err) => {
+ info!(
+ "failed to get backend while trying to rename bad chunk: {digest_str} - {err}"
+ );
+ return;
+ }
+ };
+
+ if let DatastoreBackend::S3(s3_client) = backend {
+ let suffix = format!(".{}.bad", counter);
+ let target_key = match crate::s3::object_key_from_digest_with_suffix(digest, &suffix) {
+ Ok(target_key) => target_key,
+ Err(err) => {
+ info!("could not generate target key for corrupted chunk {path:?} - {err}");
+ return;
+ }
+ };
+ let object_key = match crate::s3::object_key_from_digest(digest) {
+ Ok(object_key) => object_key,
+ Err(err) => {
+ info!("could not generate object key for corrupted chunk {path:?} - {err}");
+ return;
+ }
+ };
+ if proxmox_async::runtime::block_on(
+ s3_client.copy_object(object_key.clone(), target_key),
+ )
+ .is_ok()
+ {
+ if proxmox_async::runtime::block_on(s3_client.delete_object(object_key)).is_err() {
+ info!("failed to delete corrupt chunk on s3 backend: {digest_str}");
+ }
+ } else {
+ info!("failed to copy corrupt chunk on s3 backend: {digest_str}");
+ // Early return to leave the potentially locally cached chunk in the same state as
+ // on the object store. Verification might have failed because of connection issue
+ // after all.
+ return;
+ }
+ }
+
+ match std::fs::rename(&path, &new_path) {
+ Ok(_) => {
+ info!("corrupted chunk renamed to {:?}", &new_path);
+ }
+ Err(err) => {
+ match err.kind() {
+ std::io::ErrorKind::NotFound => { /* ignored */ }
+ _ => info!("could not rename corrupted chunk {:?} - {err}", &path),
+ }
+ }
+ };
+ }
}
diff --git a/src/backup/verify.rs b/src/backup/verify.rs
index bdbe3148b..92d3d9c49 100644
--- a/src/backup/verify.rs
+++ b/src/backup/verify.rs
@@ -76,77 +76,6 @@ impl VerifyWorker {
}
}
- fn rename_corrupted_chunk(datastore: Arc<DataStore>, digest: &[u8; 32]) {
- let (path, digest_str) = datastore.chunk_path(digest);
-
- let mut counter = 0;
- let mut new_path = path.clone();
- loop {
- new_path.set_file_name(format!("{}.{}.bad", digest_str, counter));
- if new_path.exists() && counter < 9 {
- counter += 1;
- } else {
- break;
- }
- }
-
- let backend = match datastore.backend() {
- Ok(backend) => backend,
- Err(err) => {
- info!(
- "failed to get backend while trying to rename bad chunk: {digest_str} - {err}"
- );
- return;
- }
- };
-
- if let DatastoreBackend::S3(s3_client) = backend {
- let suffix = format!(".{}.bad", counter);
- let target_key =
- match pbs_datastore::s3::object_key_from_digest_with_suffix(digest, &suffix) {
- Ok(target_key) => target_key,
- Err(err) => {
- info!("could not generate target key for corrupted chunk {path:?} - {err}");
- return;
- }
- };
- let object_key = match pbs_datastore::s3::object_key_from_digest(digest) {
- Ok(object_key) => object_key,
- Err(err) => {
- info!("could not generate object key for corrupted chunk {path:?} - {err}");
- return;
- }
- };
- if proxmox_async::runtime::block_on(
- s3_client.copy_object(object_key.clone(), target_key),
- )
- .is_ok()
- {
- if proxmox_async::runtime::block_on(s3_client.delete_object(object_key)).is_err() {
- info!("failed to delete corrupt chunk on s3 backend: {digest_str}");
- }
- } else {
- info!("failed to copy corrupt chunk on s3 backend: {digest_str}");
- // Early return to leave the potentially locally cached chunk in the same state as
- // on the object store. Verification might have failed because of connection issue
- // after all.
- return;
- }
- }
-
- match std::fs::rename(&path, &new_path) {
- Ok(_) => {
- info!("corrupted chunk renamed to {:?}", &new_path);
- }
- Err(err) => {
- match err.kind() {
- std::io::ErrorKind::NotFound => { /* ignored */ }
- _ => info!("could not rename corrupted chunk {:?} - {err}", &path),
- }
- }
- };
- }
-
fn verify_index_chunks(
&self,
index: Box<dyn IndexFile + Send>,
@@ -189,7 +118,7 @@ impl VerifyWorker {
corrupt_chunks2.lock().unwrap().insert(digest);
info!("{err}");
errors2.fetch_add(1, Ordering::SeqCst);
- Self::rename_corrupted_chunk(datastore2.clone(), &digest);
+ datastore2.rename_corrupted_chunk(&digest);
} else {
verified_chunks2.lock().unwrap().insert(digest);
}
@@ -336,7 +265,7 @@ impl VerifyWorker {
corrupt_chunks.insert(digest);
error!(message);
errors.fetch_add(1, Ordering::SeqCst);
- Self::rename_corrupted_chunk(self.datastore.clone(), &digest);
+ self.datastore.rename_corrupted_chunk(&digest);
}
fn verify_fixed_index(&self, backup_dir: &BackupDir, info: &FileInfo) -> Result<(), Error> {
--
2.47.3
_______________________________________________
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:[~2025-10-16 13:18 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-10-16 13:18 [pbs-devel] [PATCH proxmox-backup 0/6] s3 store verify: fix concurrency issues and add missing in-memory cache eviction Christian Ebner
2025-10-16 13:18 ` Christian Ebner [this message]
2025-10-16 13:18 ` [pbs-devel] [PATCH proxmox-backup 2/6] datastore: refactor rename_corrupted_chunk error handling Christian Ebner
2025-10-16 13:18 ` [pbs-devel] [PATCH proxmox-backup 3/6] verify: never hold mutex lock in async scope on corrupt chunk rename Christian Ebner
2025-10-16 13:18 ` [pbs-devel] [PATCH proxmox-backup 4/6] datastore: acquire chunk store mutex lock when renaming corrupt chunk Christian Ebner
2025-10-16 13:18 ` [pbs-devel] [PATCH proxmox-backup 5/6] datastore: verify: evict corrupt chunks from in-memory LRU cache Christian Ebner
2025-10-16 13:18 ` [pbs-devel] [PATCH proxmox-backup 6/6] verify: distinguish s3 object fetching and chunk loading error 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=20251016131819.349049-2-c.ebner@proxmox.com \
--to=c.ebner@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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.