Commit Graph

22 Commits

Author SHA1 Message Date
Harshavardhana b517c791e9
[feat]: use DSYNC for xl.meta writes and NOATIME for reads (#11615)
Instead of using O_SYNC, we are better off using O_DSYNC
instead since we are only ever interested in data to be
persisted to disk not the associated filesystem metadata.

For reads we ask customers to turn off noatime, but instead
we can proactively use O_NOATIME flag to avoid atime updates
upon reads.
2021-02-24 00:14:16 -08:00
Harshavardhana 76e2713ffe
fix: use buffers only when necessary for io.Copy() (#11229)
Use separate sync.Pool for writes/reads

Avoid passing buffers for io.CopyBuffer()
if the writer or reader implement io.WriteTo or io.ReadFrom
respectively then its useless for sync.Pool to allocate
buffers on its own since that will be completely ignored
by the io.CopyBuffer Go implementation.

Improve this wherever we see this to be optimal.

This allows us to be more efficient on memory usage.
```
   385  // copyBuffer is the actual implementation of Copy and CopyBuffer.
   386  // if buf is nil, one is allocated.
   387  func copyBuffer(dst Writer, src Reader, buf []byte) (written int64, err error) {
   388  	// If the reader has a WriteTo method, use it to do the copy.
   389  	// Avoids an allocation and a copy.
   390  	if wt, ok := src.(WriterTo); ok {
   391  		return wt.WriteTo(dst)
   392  	}
   393  	// Similarly, if the writer has a ReadFrom method, use it to do the copy.
   394  	if rt, ok := dst.(ReaderFrom); ok {
   395  		return rt.ReadFrom(src)
   396  	}
```

From readahead package
```
// WriteTo writes data to w until there's no more data to write or when an error occurs.
// The return value n is the number of bytes written.
// Any error encountered during the write is also returned.
func (a *reader) WriteTo(w io.Writer) (n int64, err error) {
	if a.err != nil {
		return 0, a.err
	}
	n = 0
	for {
		err = a.fill()
		if err != nil {
			return n, err
		}
		n2, err := w.Write(a.cur.buffer())
		a.cur.inc(n2)
		n += int64(n2)
		if err != nil {
			return n, err
		}
```
2021-01-06 09:36:55 -08:00
Harshavardhana 174f428571
add additional fdatasync before close() on writes (#9947) 2020-07-01 10:57:23 -07:00
Krishna Srinivas 94f1a1dea3
add option for O_SYNC writes for standalone FS backend (#9581) 2020-05-12 19:24:59 -07:00
Harshavardhana eff4127efd Revert "Write files in O_SYNC for fs backend to protect against machine crashes (#9434)"
This reverts commit 4843affd0e.
2020-04-27 09:22:05 -07:00
Krishna Srinivas 4843affd0e
Write files in O_SYNC for fs backend to protect against machine crashes (#9434) 2020-04-25 01:18:54 -07:00
stefan-work f001e99fcd
create the final file with mode 0666 for multipart-uploads (#9173)
NAS gateway creates non-multipart-uploads with mode 0666.
But multipart-uploads are created with a differing mode of 0644.

Both modes should be equal! Else it leads to files with different
permissions based on its file-size. This patch solves that by
using 0666 for both cases.
2020-03-20 15:32:15 -07:00
Klaus Post 8d98662633
re-implement data usage crawler to be more efficient (#9075)
Implementation overview: 

https://gist.github.com/klauspost/1801c858d5e0df391114436fdad6987b
2020-03-18 16:19:29 -07:00
Haines Chan b2044dd22f Fix typo in comments (#8704) 2019-12-27 15:51:32 +05:30
Krishna Srinivas 980bf78b4d Detect underlying disk mount/unmount (#8408) 2019-10-25 10:37:53 -07:00
Krishna Srinivas c38ada1a26 write() to disk in 4MB blocks for better performance (#7888) 2019-08-23 15:36:46 -07:00
Harshavardhana 39b3e4f9b3 Avoid using io.ReadFull() for WriteAll and CreateFile (#7676)
With these changes we are now able to peak performances
for all Write() operations across disks HDD and NVMe.

Also adds readahead for disk reads, which also increases
performance for reads by 3x.
2019-05-22 13:47:15 -07:00
kannappanr 5ecac91a55
Replace Minio refs in docs with MinIO and links (#7494) 2019-04-09 11:39:42 -07:00
Bala FA b0deea27df Refactor s3select to support parquet. (#7023)
Also handle pretty formatted JSON documents.
2019-01-08 16:53:04 -08:00
Aditya Manthramurthy 36e51d0cee Add GetObjectNInfo to object layer (#6449)
The new call combines GetObjectInfo and GetObject, and returns an
object with a ReadCloser interface.

Also adds a number of end-to-end encryption tests at the handler
level.
2018-09-20 19:22:09 -07:00
Praveen raj Mani 30d4a2cf53 s3select should honour custom record delimiter (#6419)
Allow custom delimiters like `\r\n`, `a`, `\r` etc in input csv and 
replace with `\n`.

Fixes #6403
2018-09-10 21:50:28 +05:30
Harshavardhana 4487f70f08 Revert all GetObjectNInfo related PRs (#6398)
* Revert "Encrypted reader wrapped in NewGetObjectReader should be closed (#6383)"

This reverts commit 53a0bbeb5b.

* Revert "Change SelectAPI to use new GetObjectNInfo API (#6373)"

This reverts commit 5b05df215a.

* Revert "Implement GetObjectNInfo object layer call (#6290)"

This reverts commit e6d740ce09.
2018-08-31 13:10:12 -07:00
Aditya Manthramurthy e6d740ce09 Implement GetObjectNInfo object layer call (#6290)
This combines calling GetObjectInfo and GetObject while returning a
io.ReadCloser for the object's body. This allows the two operations to
be under a single lock, fixing a race between getting object info and
reading the object body.
2018-08-27 15:28:23 +05:30
Harshavardhana 52eea7b9c1
Support SSE-C multipart source objects in CopyObject (#5603)
Current code didn't implement the logic to support
decrypting encrypted multiple parts, this PR fixes
by supporting copying encrypted multipart objects.
2018-03-02 17:24:02 -08:00
Harshavardhana 7cc678c653 Support encryption for CopyObject, GET-Range requests (#5544)
- Implement CopyObject encryption support
- Handle Range GETs for encrypted objects

Fixes #5193
2018-02-23 15:07:21 -08:00
Krishna Srinivas 3b2486ebaf Lock free multipart backend implementation for FS (#5401) 2018-01-31 13:17:24 -08:00
Andreas Auernhammer ca6b4773ed add SSE-C support for HEAD, GET, PUT (#4894)
This change adds server-side-encryption support for HEAD, GET and PUT
operations. This PR only addresses single-part PUTs and GETs without
HTTP ranges.

Further this change adds the concept of reserved object metadata which is required
to make encrypted objects tamper-proof and provide API compatibility to AWS S3.
This PR adds the following reserved metadata entries:
- X-Minio-Internal-Server-Side-Encryption-Iv          ('guarantees' tamper-proof property)
- X-Minio-Internal-Server-Side-Encryption-Kdf         (makes Key-MAC computation negotiable in future)
- X-Minio-Internal-Server-Side-Encryption-Key-Mac     (provides AWS S3 API compatibility)

The prefix `X-Minio_Internal` specifies an internal metadata entry which must not
send to clients. All client requests containing a metadata key starting with `X-Minio-Internal`
must also rejected. This is implemented by a generic-handler.

This PR implements SSE-C separated from client-side-encryption (CSE). This cannot decrypt
server-side-encrypted objects on the client-side. However, clients can encrypted the same object
with CSE and SSE-C.

This PR does not address:
 - SSE-C Copy and Copy part
 - SSE-C GET with HTTP ranges
 - SSE-C multipart PUT
 - SSE-C Gateway

Each point must be addressed in a separate PR.

Added to vendor dir:
 - x/crypto/chacha20poly1305
 - x/crypto/poly1305
 - github.com/minio/sio
2017-11-07 15:18:59 -08:00