garage/src/api/s3_get.rs

use std::sync::Arc;
use std::time::{Duration, UNIX_EPOCH};

use futures::stream::*;
use hyper::body::Bytes;
use hyper::{Body, Request, Response, StatusCode};

use garage_table::EmptyKey;

use garage_model::garage::Garage;
use garage_model::object_table::*;

use crate::error::*;

fn object_headers(
	version: &ObjectVersion,
	version_meta: &ObjectVersionMeta,
) -> http::response::Builder {
	let date = UNIX_EPOCH + Duration::from_millis(version.timestamp);
	let date_str = httpdate::fmt_http_date(date);

	let mut resp = Response::builder()
		.header(
			"Content-Type",
			version_meta.headers.content_type.to_string(),
		)
		.header("Content-Length", format!("{}", version_meta.size))
		.header("ETag", version_meta.etag.to_string())
		.header("Last-Modified", date_str)
		.header("Accept-Ranges", format!("bytes"));

	for (k, v) in version_meta.headers.other.iter() {
		resp = resp.header(k, v.to_string());
	}

	resp
}

pub async fn handle_head(
	garage: Arc<Garage>,
	bucket: &str,
	key: &str,
) -> Result<Response<Body>, Error> {
	let object = match garage
		.object_table
		.get(&bucket.to_string(), &key.to_string())
		.await?
	{
		None => return Err(Error::NotFound),
		Some(o) => o,
	};

	let version = match object
		.versions()
		.iter()
		.rev()
		.filter(|v| v.is_data())
		.next()
	{
		Some(v) => v,
		None => return Err(Error::NotFound),
	};
	let version_meta = match &version.state {
		ObjectVersionState::Complete(ObjectVersionData::Inline(meta, _)) => meta,
		ObjectVersionState::Complete(ObjectVersionData::FirstBlock(meta, _)) => meta,
		_ => unreachable!(),
	};

	let body: Body = Body::from(vec![]);
	let response = object_headers(&version, version_meta)
		.status(StatusCode::OK)
		.body(body)
		.unwrap();
	Ok(response)
}

pub async fn handle_get(
	garage: Arc<Garage>,
	req: &Request<Body>,
	bucket: &str,
	key: &str,
) -> Result<Response<Body>, Error> {
	let object = match garage
		.object_table
		.get(&bucket.to_string(), &key.to_string())
		.await?
	{
		None => return Err(Error::NotFound),
		Some(o) => o,
	};

	let last_v = match object
		.versions()
		.iter()
		.rev()
		.filter(|v| v.is_complete())
		.next()
	{
		Some(v) => v,
		None => return Err(Error::NotFound),
	};
	let last_v_data = match &last_v.state {
		ObjectVersionState::Complete(x) => x,
		_ => unreachable!(),
	};
	let last_v_meta = match last_v_data {
		ObjectVersionData::DeleteMarker => return Err(Error::NotFound),
		ObjectVersionData::Inline(meta, _) => meta,
		ObjectVersionData::FirstBlock(meta, _) => meta,
	};

	let range = match req.headers().get("range") {
		Some(range) => {
			let range_str = range.to_str()?;
			let mut ranges = http_range::HttpRange::parse(range_str, last_v_meta.size)?;
			if ranges.len() > 1 {
				return Err(Error::BadRequest(format!("Multiple ranges not supported")));
			} else {
				ranges.pop()
			}
		}
		None => None,
	};
	if let Some(range) = range {
		return handle_get_range(
			garage,
			last_v,
			last_v_data,
			last_v_meta,
			range.start,
			range.start + range.length,
		)
		.await;
	}

	let resp_builder = object_headers(&last_v, last_v_meta).status(StatusCode::OK);

	match &last_v_data {
		ObjectVersionData::DeleteMarker => unreachable!(),
		ObjectVersionData::Inline(_, bytes) => {
			let body: Body = Body::from(bytes.to_vec());
			Ok(resp_builder.body(body)?)
		}
		ObjectVersionData::FirstBlock(_, first_block_hash) => {
			let read_first_block = garage.block_manager.rpc_get_block(&first_block_hash);
			let get_next_blocks = garage.version_table.get(&last_v.uuid, &EmptyKey);

			let (first_block, version) = futures::try_join!(read_first_block, get_next_blocks)?;
			let version = match version {
				Some(v) => v,
				None => return Err(Error::NotFound),
			};

			let mut blocks = version
				.blocks()
				.iter()
				.map(|vb| (vb.hash, None))
				.collect::<Vec<_>>();
			blocks[0].1 = Some(first_block);

			let body_stream = futures::stream::iter(blocks)
				.map(move |(hash, data_opt)| {
					let garage = garage.clone();
					async move {
						if let Some(data) = data_opt {
							Ok(Bytes::from(data))
						} else {
							garage
								.block_manager
								.rpc_get_block(&hash)
								.await
								.map(Bytes::from)
						}
					}
				})
				.buffered(2);
			//let body: Body = Box::new(StreamBody::new(Box::pin(body_stream)));
			let body = hyper::body::Body::wrap_stream(body_stream);
			Ok(resp_builder.body(body)?)
		}
	}
}

pub async fn handle_get_range(
	garage: Arc<Garage>,
	version: &ObjectVersion,
	version_data: &ObjectVersionData,
	version_meta: &ObjectVersionMeta,
	begin: u64,
	end: u64,
) -> Result<Response<Body>, Error> {
	if end > version_meta.size {
		return Err(Error::BadRequest(format!("Range not included in file")));
	}

	let resp_builder = object_headers(version, version_meta)
		.header(
			"Content-Range",
			format!("bytes {}-{}/{}", begin, end, version_meta.size),
		)
		.status(StatusCode::PARTIAL_CONTENT);

	match &version_data {
		ObjectVersionData::DeleteMarker => unreachable!(),
		ObjectVersionData::Inline(_meta, bytes) => {
			if end as usize <= bytes.len() {
				let body: Body = Body::from(bytes[begin as usize..end as usize].to_vec());
				Ok(resp_builder.body(body)?)
			} else {
				None.ok_or_internal_error(
					"Requested range not present in inline bytes when it should have been",
				)
			}
		}
		ObjectVersionData::FirstBlock(_meta, _first_block_hash) => {
			let version = garage.version_table.get(&version.uuid, &EmptyKey).await?;
			let version = match version {
				Some(v) => v,
				None => return Err(Error::NotFound),
			};

			let blocks = version
				.blocks()
				.iter()
				.cloned()
				.filter(|block| block.offset + block.size > begin && block.offset < end)
				.collect::<Vec<_>>();

			let body_stream = futures::stream::iter(blocks)
				.map(move |block| {
					let garage = garage.clone();
					async move {
						let data = garage.block_manager.rpc_get_block(&block.hash).await?;
						let start_in_block = if block.offset > begin {
							0
						} else {
							begin - block.offset
						};
						let end_in_block = if block.offset + block.size < end {
							block.size
						} else {
							end - block.offset
						};
						Result::<Bytes, Error>::Ok(Bytes::from(
							data[start_in_block as usize..end_in_block as usize].to_vec(),
						))
					}
				})
				.buffered(2);
			//let body: Body = Box::new(StreamBody::new(Box::pin(body_stream)));
			let body = hyper::body::Body::wrap_stream(body_stream);
			Ok(resp_builder.body(body)?)
		}
	}
}
Some basic S3 functionnality 2020-04-24 18:47:11 +00:00			`use std::sync::Arc;`
			`use std::time::{Duration, UNIX_EPOCH};`

			`use futures::stream::*;`
			`use hyper::body::Bytes;`
Implement HTTP ranges in get 2020-05-04 13:09:23 +00:00			`use hyper::{Body, Request, Response, StatusCode};`
Some basic S3 functionnality 2020-04-24 18:47:11 +00:00
			`use garage_table::EmptyKey;`

Rename garage_core to garage_model 2020-07-07 11:59:22 +00:00			`use garage_model::garage::Garage;`
			`use garage_model::object_table::*;`
Some basic S3 functionnality 2020-04-24 18:47:11 +00:00
Refactor error management in API part 2020-11-08 14:04:30 +00:00			`use crate::error::*;`

cargo fmt 2020-07-08 15:34:37 +00:00			`fn object_headers(`
			`version: &ObjectVersion,`
			`version_meta: &ObjectVersionMeta,`
			`) -> http::response::Builder {`
Some basic S3 functionnality 2020-04-24 18:47:11 +00:00			`let date = UNIX_EPOCH + Duration::from_millis(version.timestamp);`
			`let date_str = httpdate::fmt_http_date(date);`

More headers taken into account 2020-07-09 15:04:43 +00:00			`let mut resp = Response::builder()`
cargo fmt 2020-07-08 15:34:37 +00:00			`.header(`
			`"Content-Type",`
			`version_meta.headers.content_type.to_string(),`
			`)`
Migrate S3 api to use new model 2020-07-08 15:33:24 +00:00			`.header("Content-Length", format!("{}", version_meta.size))`
			`.header("ETag", version_meta.etag.to_string())`
Some basic S3 functionnality 2020-04-24 18:47:11 +00:00			`.header("Last-Modified", date_str)`
More headers taken into account 2020-07-09 15:04:43 +00:00			`.header("Accept-Ranges", format!("bytes"));`

Validate content MD5 and SHA256 sums for PutObject and UploadPart 2020-07-15 13:31:13 +00:00			`for (k, v) in version_meta.headers.other.iter() {`
			`resp = resp.header(k, v.to_string());`
			`}`
More headers taken into account 2020-07-09 15:04:43 +00:00
Validate content MD5 and SHA256 sums for PutObject and UploadPart 2020-07-15 13:31:13 +00:00			`resp`
Some basic S3 functionnality 2020-04-24 18:47:11 +00:00			`}`

			`pub async fn handle_head(`
			`garage: Arc<Garage>,`
			`bucket: &str,`
			`key: &str,`
Update to Hyper 0.13.6 that accepts non-Sync streams in wrap_stream. Simplifies code and makes it possible to publish on crates.io 2020-07-07 15:15:53 +00:00			`) -> Result<Response<Body>, Error> {`
Some basic S3 functionnality 2020-04-24 18:47:11 +00:00			`let object = match garage`
			`.object_table`
			`.get(&bucket.to_string(), &key.to_string())`
			`.await?`
			`{`
			`None => return Err(Error::NotFound),`
			`Some(o) => o,`
			`};`

			`let version = match object`
			`.versions()`
			`.iter()`
			`.rev()`
Prepare for multipart uploads 2020-04-26 18:55:13 +00:00			`.filter(\|v\| v.is_data())`
Some basic S3 functionnality 2020-04-24 18:47:11 +00:00			`.next()`
			`{`
			`Some(v) => v,`
			`None => return Err(Error::NotFound),`
			`};`
cargo fmt 2020-07-08 15:34:37 +00:00			`let version_meta = match &version.state {`
			`ObjectVersionState::Complete(ObjectVersionData::Inline(meta, _)) => meta,`
			`ObjectVersionState::Complete(ObjectVersionData::FirstBlock(meta, _)) => meta,`
			`_ => unreachable!(),`
			`};`
Some basic S3 functionnality 2020-04-24 18:47:11 +00:00
Update to Hyper 0.13.6 that accepts non-Sync streams in wrap_stream. Simplifies code and makes it possible to publish on crates.io 2020-07-07 15:15:53 +00:00			`let body: Body = Body::from(vec![]);`
Migrate S3 api to use new model 2020-07-08 15:33:24 +00:00			`let response = object_headers(&version, version_meta)`
Some basic S3 functionnality 2020-04-24 18:47:11 +00:00			`.status(StatusCode::OK)`
			`.body(body)`
			`.unwrap();`
			`Ok(response)`
			`}`

			`pub async fn handle_get(`
			`garage: Arc<Garage>,`
Implement HTTP ranges in get 2020-05-04 13:09:23 +00:00			`req: &Request<Body>,`
Some basic S3 functionnality 2020-04-24 18:47:11 +00:00			`bucket: &str,`
			`key: &str,`
Update to Hyper 0.13.6 that accepts non-Sync streams in wrap_stream. Simplifies code and makes it possible to publish on crates.io 2020-07-07 15:15:53 +00:00			`) -> Result<Response<Body>, Error> {`
Some basic S3 functionnality 2020-04-24 18:47:11 +00:00			`let object = match garage`
			`.object_table`
			`.get(&bucket.to_string(), &key.to_string())`
			`.await?`
			`{`
			`None => return Err(Error::NotFound),`
			`Some(o) => o,`
			`};`

			`let last_v = match object`
			`.versions()`
			`.iter()`
			`.rev()`
Prepare for multipart uploads 2020-04-26 18:55:13 +00:00			`.filter(\|v\| v.is_complete())`
Some basic S3 functionnality 2020-04-24 18:47:11 +00:00			`.next()`
			`{`
			`Some(v) => v,`
			`None => return Err(Error::NotFound),`
			`};`
cargo fmt 2020-07-08 15:34:37 +00:00			`let last_v_data = match &last_v.state {`
			`ObjectVersionState::Complete(x) => x,`
			`_ => unreachable!(),`
			`};`
			`let last_v_meta = match last_v_data {`
			`ObjectVersionData::DeleteMarker => return Err(Error::NotFound),`
			`ObjectVersionData::Inline(meta, _) => meta,`
			`ObjectVersionData::FirstBlock(meta, _) => meta,`
			`};`
Some basic S3 functionnality 2020-04-24 18:47:11 +00:00
Implement HTTP ranges in get 2020-05-04 13:09:23 +00:00			`let range = match req.headers().get("range") {`
			`Some(range) => {`
Refactor error management in API part 2020-11-08 14:04:30 +00:00			`let range_str = range.to_str()?;`
			`let mut ranges = http_range::HttpRange::parse(range_str, last_v_meta.size)?;`
Implement HTTP ranges in get 2020-05-04 13:09:23 +00:00			`if ranges.len() > 1 {`
			`return Err(Error::BadRequest(format!("Multiple ranges not supported")));`
			`} else {`
			`ranges.pop()`
			`}`
			`}`
			`None => None,`
			`};`
			`if let Some(range) = range {`
cargo fmt 2020-07-08 15:34:37 +00:00			`return handle_get_range(`
			`garage,`
			`last_v,`
			`last_v_data,`
			`last_v_meta,`
			`range.start,`
			`range.start + range.length,`
			`)`
			`.await;`
Implement HTTP ranges in get 2020-05-04 13:09:23 +00:00			`}`

Migrate S3 api to use new model 2020-07-08 15:33:24 +00:00			`let resp_builder = object_headers(&last_v, last_v_meta).status(StatusCode::OK);`
Some basic S3 functionnality 2020-04-24 18:47:11 +00:00
Migrate S3 api to use new model 2020-07-08 15:33:24 +00:00			`match &last_v_data {`
			`ObjectVersionData::DeleteMarker => unreachable!(),`
			`ObjectVersionData::Inline(_, bytes) => {`
Update to Hyper 0.13.6 that accepts non-Sync streams in wrap_stream. Simplifies code and makes it possible to publish on crates.io 2020-07-07 15:15:53 +00:00			`let body: Body = Body::from(bytes.to_vec());`
Some basic S3 functionnality 2020-04-24 18:47:11 +00:00			`Ok(resp_builder.body(body)?)`
			`}`
Migrate S3 api to use new model 2020-07-08 15:33:24 +00:00			`ObjectVersionData::FirstBlock(_, first_block_hash) => {`
Some basic S3 functionnality 2020-04-24 18:47:11 +00:00			`let read_first_block = garage.block_manager.rpc_get_block(&first_block_hash);`
			`let get_next_blocks = garage.version_table.get(&last_v.uuid, &EmptyKey);`

			`let (first_block, version) = futures::try_join!(read_first_block, get_next_blocks)?;`
			`let version = match version {`
			`Some(v) => v,`
			`None => return Err(Error::NotFound),`
			`};`

			`let mut blocks = version`
			`.blocks()`
			`.iter()`
			`.map(\|vb\| (vb.hash, None))`
			`.collect::<Vec<_>>();`
			`blocks[0].1 = Some(first_block);`

			`let body_stream = futures::stream::iter(blocks)`
			`.map(move \|(hash, data_opt)\| {`
			`let garage = garage.clone();`
			`async move {`
			`if let Some(data) = data_opt {`
			`Ok(Bytes::from(data))`
			`} else {`
			`garage`
			`.block_manager`
			`.rpc_get_block(&hash)`
			`.await`
			`.map(Bytes::from)`
			`}`
			`}`
			`})`
			`.buffered(2);`
Update to Hyper 0.13.6 that accepts non-Sync streams in wrap_stream. Simplifies code and makes it possible to publish on crates.io 2020-07-07 15:15:53 +00:00			`//let body: Body = Box::new(StreamBody::new(Box::pin(body_stream)));`
Apply cargo fmt; add trace output when request signature is bad 2020-07-08 11:33:02 +00:00			`let body = hyper::body::Body::wrap_stream(body_stream);`
Some basic S3 functionnality 2020-04-24 18:47:11 +00:00			`Ok(resp_builder.body(body)?)`
			`}`
			`}`
			`}`
Implement HTTP ranges in get 2020-05-04 13:09:23 +00:00
			`pub async fn handle_get_range(`
			`garage: Arc<Garage>,`
			`version: &ObjectVersion,`
cargo fmt 2020-07-08 15:34:37 +00:00			`version_data: &ObjectVersionData,`
			`version_meta: &ObjectVersionMeta,`
Implement HTTP ranges in get 2020-05-04 13:09:23 +00:00			`begin: u64,`
			`end: u64,`
Update to Hyper 0.13.6 that accepts non-Sync streams in wrap_stream. Simplifies code and makes it possible to publish on crates.io 2020-07-07 15:15:53 +00:00			`) -> Result<Response<Body>, Error> {`
Migrate S3 api to use new model 2020-07-08 15:33:24 +00:00			`if end > version_meta.size {`
Implement HTTP ranges in get 2020-05-04 13:09:23 +00:00			`return Err(Error::BadRequest(format!("Range not included in file")));`
			`}`

Migrate S3 api to use new model 2020-07-08 15:33:24 +00:00			`let resp_builder = object_headers(version, version_meta)`
Implement HTTP ranges in get 2020-05-04 13:09:23 +00:00			`.header(`
			`"Content-Range",`
Migrate S3 api to use new model 2020-07-08 15:33:24 +00:00			`format!("bytes {}-{}/{}", begin, end, version_meta.size),`
Implement HTTP ranges in get 2020-05-04 13:09:23 +00:00			`)`
			`.status(StatusCode::PARTIAL_CONTENT);`

Migrate S3 api to use new model 2020-07-08 15:33:24 +00:00			`match &version_data {`
			`ObjectVersionData::DeleteMarker => unreachable!(),`
			`ObjectVersionData::Inline(_meta, bytes) => {`
Implement HTTP ranges in get 2020-05-04 13:09:23 +00:00			`if end as usize <= bytes.len() {`
Apply cargo fmt; add trace output when request signature is bad 2020-07-08 11:33:02 +00:00			`let body: Body = Body::from(bytes[begin as usize..end as usize].to_vec());`
Implement HTTP ranges in get 2020-05-04 13:09:23 +00:00			`Ok(resp_builder.body(body)?)`
			`} else {`
Refactor error management in API part 2020-11-08 14:04:30 +00:00			`None.ok_or_internal_error(`
			`"Requested range not present in inline bytes when it should have been",`
			`)`
Implement HTTP ranges in get 2020-05-04 13:09:23 +00:00			`}`
			`}`
Migrate S3 api to use new model 2020-07-08 15:33:24 +00:00			`ObjectVersionData::FirstBlock(_meta, _first_block_hash) => {`
Implement HTTP ranges in get 2020-05-04 13:09:23 +00:00			`let version = garage.version_table.get(&version.uuid, &EmptyKey).await?;`
			`let version = match version {`
			`Some(v) => v,`
			`None => return Err(Error::NotFound),`
			`};`

			`let blocks = version`
			`.blocks()`
			`.iter()`
			`.cloned()`
			`.filter(\|block\| block.offset + block.size > begin && block.offset < end)`
			`.collect::<Vec<_>>();`

			`let body_stream = futures::stream::iter(blocks)`
			`.map(move \|block\| {`
			`let garage = garage.clone();`
			`async move {`
			`let data = garage.block_manager.rpc_get_block(&block.hash).await?;`
			`let start_in_block = if block.offset > begin {`
			`0`
			`} else {`
			`begin - block.offset`
			`};`
			`let end_in_block = if block.offset + block.size < end {`
			`block.size`
			`} else {`
			`end - block.offset`
			`};`
Apply cargo fmt; add trace output when request signature is bad 2020-07-08 11:33:02 +00:00			`Result::<Bytes, Error>::Ok(Bytes::from(`
Implement HTTP ranges in get 2020-05-04 13:09:23 +00:00			`data[start_in_block as usize..end_in_block as usize].to_vec(),`
			`))`
			`}`
			`})`
			`.buffered(2);`
Update to Hyper 0.13.6 that accepts non-Sync streams in wrap_stream. Simplifies code and makes it possible to publish on crates.io 2020-07-07 15:15:53 +00:00			`//let body: Body = Box::new(StreamBody::new(Box::pin(body_stream)));`
Apply cargo fmt; add trace output when request signature is bad 2020-07-08 11:33:02 +00:00			`let body = hyper::body::Body::wrap_stream(body_stream);`
Implement HTTP ranges in get 2020-05-04 13:09:23 +00:00			`Ok(resp_builder.body(body)?)`
			`}`
			`}`
			`}`