Commit Graph

290 Commits

Author SHA1 Message Date
Harshavardhana 653ea50a59 Erasure technique is now uint8, handle it appropriately 2015-03-24 19:12:52 -07:00
Anand Babu (AB) Periasamy fa9b1f341b erasure encode janitor duty 2015-03-24 18:47:53 -07:00
Harshavardhana 6a87ed001c Handle errors properly during erasure Decoding, also populate technique and verify 2015-03-24 18:14:13 -07:00
Harshavardhana 0474439b43 Code restructuring, assigning proper subsystems to each
- Drivers contain
    * donut/*
    * file/*
    * memory/*

 - Storage format contains
    * donut/*
      - GetObject() --> renamed to GetObjectReader()
      - Deleted stale objectwriter.go, renamed donutwriter.go to object_writer.go

Simplifying, and documenting codebase further
2015-03-24 06:47:10 -07:00
Frederick F. Kautz IV 3c4012f1e7 Adding donut backend and setting as default 2015-03-23 20:07:22 -07:00
Harshavardhana 7c1c4b8a02 Merge pull request #351 from harshavardhana/pr_out_implement_log_package_as_drop_in_replacement_for_handling_debug_log_level 2015-03-23 18:55:14 -07:00
Harshavardhana 914962bd93 Implement log package as drop-in replacement for handling Debug log-level 2015-03-23 18:18:11 -07:00
Frederick F. Kautz IV f91f5e59ee Running gofmt manually, auto gofmt didn't simplify code 2015-03-23 12:26:41 -07:00
Frederick F. Kautz IV 56a298fd6a Refactoring method variable names for erasureWriter struct 2015-03-23 12:25:20 -07:00
Frederick F. Kautz IV d57e9d6c18 Adding sys. as reserved in erasureWriter 2015-03-23 11:56:24 -07:00
Frederick F. Kautz IV b2f47ef446 Switching content hashing from sha512 to md5 2015-03-23 11:46:19 -07:00
Frederick F. Kautz IV b2c3172095 Adding sha512 support for pre-encoded data and verification on decode 2015-03-22 20:11:25 -07:00
Frederick F. Kautz IV befac7d047 Exorcising donutbox 2015-03-22 19:55:15 -07:00
Frederick F. Kautz IV eb8005cc65 Reading correct block size in erasure coded data 2015-03-22 18:49:30 -07:00
Harshavardhana 3929135c34 Make sure we use O_EXCL with O_CREATE to make sure we don't trip over existing file 2015-03-22 15:55:40 -07:00
Harshavardhana 2fa9320df2 De-couple donut into smaller files, useful for ease in external integration 2015-03-22 15:43:37 -07:00
Frederick F. Kautz IV 55f3659bb3 Removing log tracers 2015-03-22 15:39:50 -07:00
Frederick F. Kautz IV 03f51eac29 Adding list objects to donut 2015-03-22 15:33:52 -07:00
Harshavardhana 4b9fbd3b3a DonutfileWriter should return value of file.Close() 2015-03-22 14:51:18 -07:00
Frederick F. Kautz IV ced6edc9fe Minor change, removing unnecessary commented method 2015-03-22 12:44:45 -07:00
Frederick F. Kautz IV 99cf484488 Adding GetObjectMetadata to donut 2015-03-22 12:36:51 -07:00
Frederick F. Kautz IV 703af6c3ce Removing gob from donut driver 2015-03-22 12:17:12 -07:00
Harshavardhana 95affcb119 Fix build issues for new DonutDriver() 2015-03-22 02:53:21 -07:00
Frederick F. Kautz IV 933d44c656 New donut driver with local file encoding 2015-03-22 00:40:21 -07:00
Harshavardhana 2b0b5f014a Update license header Mini should be Minimalist really 2015-03-19 14:35:50 -07:00
Harshavardhana c3602967ab Fix bucketpolicy to handle s3 and minio policy names separately.
Also fix an apparent bug in GetBucketPolicy{}
2015-03-19 12:22:14 -07:00
Frederick F. Kautz IV 2c69433588 Adding interfaces for donut driver 2015-03-18 14:27:10 -07:00
Frederick F. Kautz IV bf86e78714 Renaming files to match 2015-03-18 12:26:44 -07:00
Frederick F. Kautz IV 1c9e550ee6 Renaming encoded driver to donutstorage 2015-03-18 12:25:37 -07:00
Frederick F. Kautz IV 982a52e0d4 Merge pull request #320 from fkautz/pr_out_removing_block_size_from_donut_writer_no_longer_necessary 2015-03-17 14:25:08 -07:00
Frederick F. Kautz IV d6be8d3517 Removing block size from donut writer, no longer necessary 2015-03-17 14:24:59 -07:00
Harshavardhana 74b3d092f2 Now MinioAPI handles Content-MD5 set during putObject()
- This change also facilitates proper error response
    in case of md5 mismatch or corruption
  - TODO a test function needs to be implemented
2015-03-17 13:32:13 -07:00
Harshavardhana 42006c2ab0 Merge pull request #318 from harshavardhana/pr_out_remove_contenttype_pseudo_func_ 2015-03-17 00:47:46 -07:00
Harshavardhana dccf32e5c4 Remove contenttype pseudo func() 2015-03-17 00:47:22 -07:00
Frederick F. Kautz IV 90b337a3e8 Fixing order of input and layout in time.Parse 2015-03-17 00:47:17 -07:00
Frederick F. Kautz IV 25725c442a Moving to more accurate RFC3339 2015-03-17 00:45:50 -07:00
Frederick F. Kautz IV e36fa95f1f Moving from iso8601 to RFC3339 2015-03-17 00:42:08 -07:00
Frederick F. Kautz IV 598eb5192e Encoded now passes tests, enabling. 2015-03-17 00:33:49 -07:00
Harshavardhana 6d71935ea5 Encoded storage type now feature complete 2015-03-16 23:42:45 -07:00
Frederick F. Kautz IV 34cf6682fe list objects in buckets 2015-03-16 20:43:05 -07:00
Harshavardhana b495f15178 Make blockSize as constant 2015-03-16 18:20:01 -07:00
Frederick F. Kautz IV 583c32c774 Metadata now stores hex md5, not base64. Minio API now converts md5 hex to base64. 2015-03-16 18:03:22 -07:00
Frederick F. Kautz IV 10ae089c46 Better handling of metadata in encoded 2015-03-16 17:40:59 -07:00
Harshavardhana 15f68972a5 Some necessary cleanup
- Rename 'fs' as 'file' for brevity
  - Rename 'inmemory' as 'memory' for brevity
  - Change everywhere else
2015-03-16 11:46:16 -07:00
Harshavardhana bf04d125fd Merge pull request #304 from harshavardhana/pr_out_move_from_fmt_errorf_to_errors_new 2015-03-15 19:42:23 -07:00
Harshavardhana 596eeb097e Move from fmt.Errorf to errors.New
In many cases fmt.Errorf is good enough, but since
error is an interface, you can use arbitrary data
structures as error values, to allow callers to
inspect the details of the error.
2015-03-15 19:41:16 -07:00
Frederick F. Kautz IV a589ddaddf Decoding and returning objects now works 2015-03-15 19:02:00 -07:00
Frederick F. Kautz IV 682bf085c4 Metadata and writing the object are now both committed together. 2015-03-15 17:42:33 -07:00
Frederick F. Kautz IV 96d0a089b8 Adding create bucket to encoded 2015-03-15 12:28:24 -07:00
Harshavardhana 3f95c55067 Skip uinimplemented tests and some cleanup 2015-03-15 01:23:57 -07:00
Frederick F. Kautz IV 3e453a394a Working in memory donut driver 2015-03-15 00:41:06 -07:00
Frederick F. Kautz IV 85da9fc627 Adding start of inmem donut 2015-03-14 18:02:27 -07:00
Frederick F. Kautz IV 54472c18fe Renaming singledisk to encoded 2015-03-14 11:37:43 -07:00
Frederick F. Kautz IV 348c4d9a88 Removing old donut format 2015-03-14 11:36:19 -07:00
Frederick F. Kautz IV 2f4b92eb8d Minor refactor to prepare for file version of donut 2015-03-14 11:34:45 -07:00
Harshavardhana f7fd526b28 Cosmetic changes 2015-03-12 02:11:04 -07:00
Frederick F. Kautz IV 94fe03cfc1 More refactoring to meet new donut spec 2015-03-12 01:52:14 -07:00
Harshavardhana 67888cbaa9 Restructure and use human readable names 2015-03-12 00:36:33 -07:00
Harshavardhana e21020b337 Merge pull request #280 from harshavardhana/pr_out_http_range_support_wip
Get object range request added
2015-03-12 00:11:38 -07:00
Harshavardhana a42a48a455 Get object range request added 2015-03-12 00:07:27 -07:00
Frederick F. Kautz IV 24d5018037 Rename StoreBucket to CreateBucket 2015-03-11 20:14:40 -07:00
Frederick F. Kautz IV 8c6aa45458 Adding copyright 2015-03-11 19:59:20 -07:00
Frederick F. Kautz IV 7971b66e60 Adding object donut type 2015-03-11 19:56:18 -07:00
Frederick F. Kautz IV ec9b20a94f Adding stubes for new single disk driver 2015-03-11 13:36:21 -07:00
Harshavardhana 41653774fc Simplify fs codebase, split them into separate files and more commenting 2015-03-08 18:01:48 -07:00
Harshavardhana 01e98eb326 Make BucketPolicy strictly typed 2015-03-08 15:57:20 -07:00
Harshavardhana 8866b2cc5f Add delimiter and prefix tests, also add inmemory to support delimiters 2015-03-08 14:51:31 -07:00
Harshavardhana 7a01db8005 Strict checks for delimiter and prefix if absent or present 2015-03-08 13:01:39 -07:00
Harshavardhana 9e89af0e3c Remove redundant uint32 2015-03-07 19:15:51 -08:00
Harshavardhana 9553eb6b4a Merge pull request #269 from harshavardhana/pr_out_get_object_metadata_needs_to_filter_based_of_prefix_when_delimiter_is_specified 2015-03-07 19:14:41 -08:00
Harshavardhana f8e2b75e23 Get object metadata needs to filter based of Prefix when delimiter is specified 2015-03-07 19:13:15 -08:00
Frederick F. Kautz IV b39eac36ea Merge pull request #268 from fkautz/pr_out_fragment_reader_implemented 2015-03-07 18:04:29 -08:00
Frederick F. Kautz IV 679c6ace3a Fragment reader implemented 2015-03-07 18:04:20 -08:00
Harshavardhana df69038b04 Merge pull request #267 from harshavardhana/pr_out_add_md5_support_for_fsstorage_also_re_purpose_getobjectmetadata_for_listobjects 2015-03-07 17:16:00 -08:00
Harshavardhana 4751ed0e30 Add MD5 support for fsStorage, also re-purpose GetObjectMetadata for ListObjects 2015-03-07 17:15:00 -08:00
Frederick F. Kautz IV ca1a4b616c Adding read header for erasure 2015-03-07 15:53:41 -08:00
Frederick F. Kautz IV 00381c887a Adding erasure package to wrap versioned erasure. Ensures we always write latest version. 2015-03-07 13:04:02 -08:00
Harshavardhana d992bccd9a Move pkg/storage/erasure to pkg/encoding/erasure - and other cleanups 2015-03-07 00:16:48 -08:00
Frederick F. Kautz IV 856e0100c0 Erasure layer now writes using new technique 2015-03-06 16:37:44 -08:00
Harshavardhana e5af8a3f5d Further fixes -
- All test files have been renamed to their respective <package>_test name,
    this is done in accordance with
      - https://github.com/golang/go/wiki/CodeReviewComments#import-dot
        imports are largely used in testing, but to avoid namespace collision
        and circular dependencies

  - Never use _* in package names other than "_test" change fragment_v1 to expose
    fragment just like 'gopkg.in/check.v1'
2015-03-06 02:04:21 -08:00
Harshavardhana c00d1461b9 Golint cleanup from top level 2015-03-05 23:37:37 -08:00
Harshavardhana 540723d7ae Golint cleanup pkg/api 2015-03-05 21:09:19 -08:00
Harshavardhana 76e601b26b Golint cleanup pkg/storage 2015-03-05 21:09:19 -08:00
Frederick F. Kautz IV 1186c47603 Fragment now implements new version 2015-03-05 20:56:02 -08:00
Frederick F. Kautz IV 665af44c70 Renaming donut packages to match layout 2015-03-05 19:44:01 -08:00
Harshavardhana f2e238199b Handle individual objects too 2015-03-04 01:54:35 -08:00
Harshavardhana e400e85639 Add proper path prefix to be delimited 2015-03-04 00:54:39 -08:00
Harshavardhana cc46475c86 Pathprefix is filtered out properly and delimited as well 2015-03-03 19:45:54 -08:00
Harshavardhana 3a3c8645fc Update comments across the codebase 2015-03-03 02:39:38 -08:00
Harshavardhana 137584d658 Add comments 2015-03-03 01:25:45 -08:00
Harshavardhana c3ad0906e0 Add deadcode code which recursivley goes into all directories and verifies dangling variables. 2015-03-02 18:44:20 -08:00
Frederick F. Kautz IV 32f8c17174 Renaming AppendUstr to AppendUniqStr 2015-03-01 17:06:00 -08:00
Frederick F. Kautz IV 6f300c9ef0 Adding test at byte level 2015-03-01 15:13:09 -08:00
Frederick F. Kautz IV 0b08312ddd Specifying more exact types for data 2015-03-01 14:43:11 -08:00
Harshavardhana 9f43659135 Add tests, add NewHeader() for populating Header and ValidateHeader() 2015-03-01 13:19:06 -08:00
Anand Babu (AB) Periasamy 6515678e87 Merge pull request #236 from fkautz/pr_out_adding_data_writer
Adding data writer
2015-03-01 11:36:35 -08:00
Frederick F. Kautz IV 7414dc9b75 Adding data writer 2015-03-01 11:32:30 -08:00
Harshavardhana eb9a4a83ac Move AppendUstr(), AppendUint() into helpers 2015-03-01 11:19:03 -08:00
Frederick F. Kautz IV be12a6b3d9 Changing K and M in data.go to match new datatype in erasure package 2015-03-01 01:28:33 -08:00
Harshavardhana 33601ff567 Merge pull request #233 from harshavardhana/pr_out_make_k_m_to_be_uint8_and_technique_becomes_its_own_type 2015-03-01 01:26:42 -08:00
Harshavardhana b77c62e25d Make K,M to be uint8 and Technique becomes its own type 2015-03-01 01:25:28 -08:00
Frederick F. Kautz IV bb0441917c Making changes to data based on discussion 2015-03-01 01:16:22 -08:00
Harshavardhana 2f04a1299e Avoid allocating 1gig memory, avoid such mistakes 2015-03-01 00:43:12 -08:00
Frederick F. Kautz IV 2916b1278e Adding data.go, does nothing for now. 2015-03-01 00:27:52 -08:00
Frederick F. Kautz IV b4ef522a9b Donut frame is now its own package 2015-03-01 00:02:14 -08:00
Frederick F. Kautz IV 6c2428833a Renaming Write to WriteFrame 2015-02-28 23:52:56 -08:00
Harshavardhana 6b397844b9 Add donut benchmark for 64MB, 128MB, 256MB, 512MB
Current average values

~~~
> minio/pkg/storage/donut/v1 *> go test -bench .
OK: 2 passed
PASS
BenchmarkDonut64M              5         266249634 ns/op         252.05 MB/s
BenchmarkDonut128M             3         436507394 ns/op         307.48 MB/s
BenchmarkDonut256M             2         836116359 ns/op         321.05 MB/s
BenchmarkDonut512M             1        1676738951 ns/op         320.19 MB/s
ok      github.com/minio-io/minio/pkg/storage/donut/v1  8.430s
> minio/pkg/storage/donut/v1 *>
~~~
2015-02-28 17:05:46 -08:00
Harshavardhana 65f22083ca donut_gen works with new Donut Frame 2015-02-28 16:30:18 -08:00
Frederick F. Kautz IV ddc7cf835e Donut now tests amount written should match expected amount to write 2015-02-28 16:09:52 -08:00
Frederick F. Kautz IV d3d1205eb1 Updating comment describing data type 2015-02-28 15:59:52 -08:00
Frederick F. Kautz IV 58a04ee831 New donut frame implemented 2015-02-28 15:54:13 -08:00
Harshavardhana 0c2d58bc6d Implement proper delimiter and prefix handling
With this change Minio server now responds with, delimited
'object names' in conjunction with prefix filtering

~~~
<ListBucketResult>
  <Name>example-bucket</Name>
  <Prefix></Prefix>
  <Marker></Marker>
  <MaxKeys>1000</MaxKeys>
  <Delimiter>/</Delimiter>
  <IsTruncated>false</IsTruncated>
  <Contents>
    <Key>sample.html</Key>
    <LastModified>2011-02-26T01:56:20.000Z</LastModified>
    <ETag>example-bucket#sample.html</ETag>
    <Size>142863</Size>
    <Owner>
      <ID>minio</ID>
      <DisplayName>minio</DisplayName>
    </Owner>
    <StorageClass>STANDARD</StorageClass>
  </Contents>
  <CommonPrefixes>
    <Prefix>photos/</Prefix>
  </CommonPrefixes>
</ListBucketResult>
~~~

~~~
<ListBucketResult>
  <Name>example-bucket</Name>
  <Prefix>photos/2006/</Prefix>
  <Marker></Marker>
  <MaxKeys>1000</MaxKeys>
  <Delimiter>/</Delimiter>
  <IsTruncated>false</IsTruncated>

  <CommonPrefixes>
    <Prefix>photos/2006/feb/</Prefix>
  </CommonPrefixes>
  <CommonPrefixes>
    <Prefix>photos/2006/jan/</Prefix>
  </CommonPrefixes>
</ListBucketResult>
~~~
2015-02-28 14:47:19 -08:00
Harshavardhana d5e5a3b3e6 Use validString 2015-02-27 19:49:18 -08:00
Harshavardhana 6ebb48b4ea Add object name validation 2015-02-27 19:42:04 -08:00
Harshavardhana 53669a0854 Implement delimiter, path prefix 2015-02-27 16:23:05 -08:00
Harshavardhana 3f8b3462e6 Remove donut_gen binary and update crc32c function 2015-02-27 14:41:53 -08:00
Frederick F. Kautz IV 2e37791681 Merge pull request #212 from fkautz/pr_out_adding_comparison_for_footer_s_inim 2015-02-27 12:40:10 -08:00
Frederick F. Kautz IV 1f7a3c13a5 Adding comparison for footer's INIM 2015-02-27 12:39:52 -08:00
Frederick F. Kautz IV 2ca62f13a5 Moving donut_gen to its own dir and renaming it as donut_gen_v1 2015-02-27 12:39:14 -08:00
Frederick F. Kautz IV be768d01a2 Seek to beginning for tempbuffer 2015-02-27 12:18:27 -08:00
Frederick F. Kautz IV 5c52da6ebe Adding crc32c to write 2015-02-26 23:04:07 -08:00
Frederick F. Kautz IV ef442289da Making donut use io.ReadWriteSeek instead of io.Writer 2015-02-23 20:27:25 -08:00
Harshavardhana f4b5519451 Update erasure documentation 2015-02-23 13:39:16 -08:00
Harshavardhana 51e80eaa6d Add domain and subdomain support for MinioAPI
This change brings in domain and subdomain support

   - ./minio --domain "yourminiodomain.com"

This change brings in a much needed feature by keeping
bucketnames as part of your 'DNS' name.

All your existing applications can be migrated off from s3 to
Minio without little to no modifications.

NOTE: Setting up DNS for your `buckets` is out of scope of this feature
2015-02-23 02:25:01 -08:00
Harshavardhana 3151f9b6c3 Add proper entries for GobHeader, Data copy.
This commit also adds 'donut_gen.go' for generating sample donut file
2015-02-21 01:44:27 -08:00
Frederick F. Kautz IV 8129d8fd0b Writing data out with error handling 2015-02-20 23:24:01 -08:00
Frederick F. Kautz IV 98ddbd840e Moving donut code to v1 in donut package 2015-02-20 23:07:43 -08:00
Frederick F. Kautz IV 9833647fb7 Fixing magic number for INIM 2015-02-20 22:42:58 -08:00
Anand Babu (AB) Periasamy ec36a598c9 Magic as numbers are faster to encode and compare 2015-02-20 18:20:21 -08:00
Harshavardhana c4f0aad9b5 Add donut tests, just for verification 2015-02-18 19:46:14 -08:00
Frederick F. Kautz IV f7f34d08e8 Donut structure refined 2015-02-18 17:18:33 -08:00
Frederick F. Kautz IV ab80c0ce12 Merge pull request #180 from abperiasamy/donut 2015-02-18 15:57:16 -08:00
Anand Babu (AB) Periasamy 9e304b53de donut spec v1 prototype 2015-02-18 15:52:12 -08:00
Harshavardhana fab954f13f Run govet and fix 2015-02-18 15:33:55 -08:00
Frederick F. Kautz IV 05ac09bf69 Adding header version 2015-02-18 10:21:02 -08:00
Harshavardhana a3e5121f7b Make channels more aware of errors and shutdown on error - some cleanup 2015-02-18 00:24:19 -08:00
Frederick F. Kautz IV fe9a5c14a0 Adding scaffolding for writing donut 2015-02-17 17:47:10 -08:00
Harshavardhana f04beab7c2 Skip policy files 2015-02-16 18:37:09 -08:00
Harshavardhana eeae64935e Implement bucket policy handler and with galore of cleanup 2015-02-16 18:20:40 -08:00
Harshavardhana 966cdd4f05 Reply in xml from validate and ignore handlers 2015-02-11 22:00:45 -08:00
Harshavardhana ac4f07906c Implement S3 Style ErrorCodes and Response
This patchset also brings in lot of cleanup in terms of minioapi codebase
2015-02-11 21:27:28 -08:00
Harshavardhana 680848bdcb Add license header 2015-02-08 03:00:32 -08:00
Frederick F. Kautz IV 288bc3fcc7 Adding content type to backend storage drivers 2015-02-04 17:32:40 -08:00
Frederick F. Kautz IV 7152712f34 Accessing dir as object should return ObjectNotFound 2015-02-03 16:45:00 -08:00
Frederick F. Kautz IV a092b44cf1 Adding test to ensure list objects returns an error for non existant buckets to match previous bug fix and fix for inmemory 2015-02-01 10:48:31 -08:00
Harshavardhana 5d197dd063 Reply back errors properly 2015-02-01 03:10:28 -08:00
Frederick F. Kautz IV 54026dbc78 Setting order when retrieving buckets 2015-02-01 00:16:25 -08:00
Frederick F. Kautz IV 7ca87e06aa Adding list bucket tests and fixing inmemory behavior. 2015-01-29 11:23:14 -08:00
Harshavardhana 82baccb8bf Prefix if empty skip it before IsValidBucket 2015-01-29 01:07:37 -08:00
Frederick F. Kautz IV ca7b1a7fc6 Fix build 2015-01-28 14:00:08 -08:00