Store content files into the S3

file-store/src/lib.rs

@@ -22,20 +22,6 @@ pub enum Error {
 
 pub type Result<T> = std::result::Result<T, Error>;
 
-impl Deref for File {
-    type Target = NamedTempFile;
-
-    fn deref(&self) -> &Self::Target {
-        &self.file
-    }
-}
-
-impl DerefMut for File {
-    fn deref_mut(&mut self) -> &mut Self::Target {
-        &mut self.file
-    }
-}
-
 #[derive(Clone, Debug)]
 pub struct FileStore {
     path: PathBuf,
@@ -146,6 +132,20 @@ impl File {
     }
 }
 
+impl Deref for File {
+    type Target = NamedTempFile;
+
+    fn deref(&self) -> &Self::Target {
+        &self.file
+    }
+}
+
+impl DerefMut for File {
+    fn deref_mut(&mut self) -> &mut Self::Target {
+        &mut self.file
+    }
+}
+
 #[cfg(test)]
 mod test {
     use std::io::Write;

index-scheduler/src/batch.rs

@@ -198,6 +198,35 @@ impl Batch {
             | IndexDocumentDeletionByFilter { index_uid, .. } => Some(index_uid),
         }
     }
+
+    /// Return the content fields uuids associated with this batch.
+    pub fn content_uuids(&self) -> Vec<Uuid> {
+        match self {
+            Batch::TaskCancelation { .. }
+            | Batch::TaskDeletion(_)
+            | Batch::Dump(_)
+            | Batch::IndexCreation { .. }
+            | Batch::IndexDocumentDeletionByFilter { .. }
+            | Batch::IndexUpdate { .. }
+            | Batch::SnapshotCreation(_)
+            | Batch::IndexDeletion { .. }
+            | Batch::IndexSwap { .. } => vec![],
+            Batch::IndexOperation { op, .. } => match op {
+                IndexOperation::DocumentOperation { operations, .. } => operations
+                    .iter()
+                    .flat_map(|op| match op {
+                        DocumentOperation::Add(uuid) => Some(*uuid),
+                        DocumentOperation::Delete(_) => None,
+                    })
+                    .collect(),
+                IndexOperation::DocumentDeletion { .. }
+                | IndexOperation::Settings { .. }
+                | IndexOperation::DocumentClear { .. }
+                | IndexOperation::SettingsAndDocumentOperation { .. }
+                | IndexOperation::DocumentClearAndSetting { .. } => vec![],
+            },
+        }
+    }
 }
 
 impl IndexOperation {

index-scheduler/src/lib.rs

@@ -35,7 +35,6 @@ pub type TaskId = u32;
 use std::collections::{BTreeMap, HashMap};
 use std::fs::File;
 use std::ops::{Bound, RangeBounds};
-use std::os::fd::AsRawFd;
 use std::path::{Path, PathBuf};
 use std::sync::atomic::AtomicBool;
 use std::sync::atomic::Ordering::Relaxed;
@@ -59,7 +58,7 @@ use parking_lot::{MappedRwLockReadGuard, RwLock, RwLockReadGuard};
 use roaring::RoaringBitmap;
 use s3::Bucket;
 use synchronoise::SignalEvent;
-use tempfile::{NamedTempFile, TempDir};
+use tempfile::TempDir;
 use time::format_description::well_known::Rfc3339;
 use time::OffsetDateTime;
 use utils::{filter_out_references_to_newer_tasks, keep_tasks_within_datetimes, map_bound};
@@ -281,7 +280,7 @@ pub struct IndexSchedulerOptions {
     /// zookeeper client
     pub zookeeper: Option<Arc<ZooKeeper>>,
     /// S3 bucket
-    pub s3: Option<Bucket>,
+    pub s3: Option<Arc<Bucket>>,
 }
 
 /// Structure which holds meilisearch's indexes and schedules the tasks
@@ -290,6 +289,7 @@ pub struct IndexSchedulerOptions {
 pub struct IndexScheduler {
     inner: Arc<RwLock<Option<IndexSchedulerInner>>>,
     zookeeper: Option<Arc<ZooKeeper>>,
+    pub s3: Option<Arc<Bucket>>,
     wake_up: Arc<SignalEvent>,
 }
 
@@ -300,6 +300,7 @@ impl IndexScheduler {
         #[cfg(test)] test_breakpoint_sdr: crossbeam::channel::Sender<(Breakpoint, bool)>,
         #[cfg(test)] planned_failures: Vec<(usize, tests::FailureLocation)>,
     ) -> Result<Self> {
+        let s3 = options.s3.clone();
         let inner = IndexSchedulerInner::new(
             options,
             #[cfg(test)]
@@ -334,6 +335,7 @@ impl IndexScheduler {
 
         let this = IndexScheduler {
             zookeeper: inner.zookeeper.clone(),
+            s3,
             wake_up: inner.wake_up.clone(),
             inner: Arc::new(RwLock::new(Some(inner))),
         };
@@ -1242,6 +1244,16 @@ impl IndexSchedulerInner {
                 Some(batch) => batch,
                 None => return Ok(TickOutcome::WaitForSignal),
             };
+
+        if let Some(s3) = &self.options.s3 {
+            for uuid in batch.content_uuids() {
+                // TODO use a real UUIDv4
+                let (_, file) = self.file_store.new_update_with_uuid(uuid.as_u128())?;
+                s3.get_object_to_writer(&format!("/update-files/{}", uuid), &mut &*file).unwrap();
+                file.persist()?;
+            }
+        }
+
         let index_uid = batch.index_uid().map(ToOwned::to_owned);
         drop(rtxn);
 

meilisearch/src/lib.rs

@@ -245,13 +245,15 @@ fn open_or_create_database_unchecked(
         instance_features,
         zookeeper: zookeeper.clone(),
         s3: opt.s3_url.as_ref().map(|url| {
-            Bucket::new(
-                "test-rust-s3",
-                Region::Custom { region: "eu-central-1".to_owned(), endpoint: url.clone() },
-                Credentials::default().unwrap(),
+            Arc::new(
+                Bucket::new(
+                    "test-rust-s3",
+                    Region::Custom { region: "eu-central-1".to_owned(), endpoint: url.clone() },
+                    Credentials::default().unwrap(),
+                )
+                .unwrap()
+                .with_path_style(),
             )
-            .unwrap()
-            .with_path_style()
         }),
     }))
     .map_err(anyhow::Error::from);

meilisearch/src/routes/indexes/documents.rs

@@ -1,4 +1,4 @@
-use std::io::ErrorKind;
+use std::io::{BufReader, ErrorKind};
 
 use actix_web::http::header::CONTENT_TYPE;
 use actix_web::web::Data;
@@ -400,6 +400,7 @@ async fn document_addition(
     }
 
     let read_file = buffer.into_inner().into_std().await;
+    let s3 = index_scheduler.s3.clone();
     let documents_count = tokio::task::spawn_blocking(move || {
         let documents_count = match format {
             PayloadType::Json => read_json(&read_file, update_file.as_file_mut())?,
@@ -408,8 +409,15 @@ async fn document_addition(
             }
             PayloadType::Ndjson => read_ndjson(&read_file, update_file.as_file_mut())?,
         };
+
+        if let Some(s3) = s3 {
+            let mut reader = BufReader::new(&*update_file);
+            s3.put_object_stream(&mut reader, format!("/update-files/{}", uuid)).unwrap();
+        }
+
         // we NEED to persist the file here because we moved the `udpate_file` in another task.
         update_file.persist()?;
+
         Ok(documents_count)
     })
     .await;