maxmustermann/minio
1
0
Fork 0
Code Issues Pull requests Projects Releases Wiki Activity
2813 commits 18 branches 277 tags 146 MiB
Commit graph

2813 commits

This branch
This branch
All branches
Author SHA1 Message Date
Harshavardhana 86e5d71519 erasure: MakeVol, DeleteVol and StatVol should hold locks. (#1597) 
Since there is a good amount of overlap, each code has to lock
properly for the operation they are going to perform.

- MakeVol create vols in a routine on all disks, hold locks.
- DeleteVol delete vols in a routine on all disks, hold locks.
- StatVol stat vols in a routine on all disks, hold locks.

Fixes #1588
 2016-05-11 12:54:21 -07:00
Harshavardhana 72748d2073 erasure: healVolume err should be different from shadowed version. (#1590) 
Multiple go-routines updating the same shadowed variable can
cause a data race, avoid it by using its own err variable.

Fixes #1589
 2016-05-11 01:36:09 -07:00
Harshavardhana 49141eb3e4 http: Remove minhttp package and use standard Golang http. (#1587) 
The functionality provided by minhttp will be implemented
cleanly through our own APIs. Since we are not going to
send SIGUSR2 and manage configuration in that manner, it
doesn't make sense to use minhttp.

Fixes #1586
 2016-05-10 18:03:00 -07:00
Harshavardhana d1fa1d9352 Remove binary files from previous commit. 2016-05-10 15:49:17 -07:00
karthic rao 26e2c4bf4d Replacing fastsha256 with crypto/sha256 package from golang standard package (#1584) 2016-05-10 14:20:11 -07:00
Krishna Srinivas b044336329 XL/GetObject: If the offset does not fall in the first "dataBlock" it gives incorrect data. (#1583) 
Fixes #1582
 2016-05-10 11:38:49 -07:00
Krishna Srinivas e99cb05516 XL/GetObject: offset should be reset to 0 after reading first part. (#1580) (#1581) 2016-05-10 10:38:12 -07:00
Krishna Srinivas 409e09c1e5 XL/Selfheal: skip reading from disk if ReadFile had returned error. (#1575) 2016-05-10 01:24:58 -07:00
Krishna Srinivas c314a98c1a XL/list: fix panic on list when a disk is down. (#1562) 2016-05-10 00:35:29 -07:00
Harshavardhana 5f0ca64346 erasure: listOnlineDisks should return errFileNotFound for errReadQuorum. (#1573) 
Fixes #1571
 2016-05-10 00:10:34 -07:00
Harshavardhana 0db3218d5d xl: getPartsMetadata fetch parts and decode in go-routine. (#1569) 
Ref #1516
 2016-05-09 23:51:05 -07:00
Harshavardhana eec41c369c posix: Return diskNotFound error rather than errVolumeNotFound (#1568) 
Fixes #1559
 2016-05-09 18:57:39 -07:00
Harshavardhana b66c3bf35e server: Enable server profiling as needed. (#1565) 2016-05-09 16:18:56 -07:00
Harshavardhana f733120d3d xl: CompleteMultipartUpload make sure to delete uploads.json (#1539) 
Fixes #1537

Ref #1540 - for missing functionality in this patch.
 2016-05-09 12:09:48 -07:00
Krishna Srinivas 6627388dc3 posix: remove dead code related to posix reserved suffixes. (#1555) 2016-05-09 11:40:30 -07:00
Harshavardhana 9d41414fb5 posix: reserved files should be filtered out at posix not object layer. (#1554) 2016-05-09 02:53:08 -07:00
Harshavardhana 722abe2d0f xl/fs: pathJoin now takes variadic inputs. (#1550) 
Retains slash for the last element.

Fixes #1546
 2016-05-09 00:46:54 -07:00
Krishna Srinivas 04a5b25929 Multipart: Minimum part size limit does not apply to the last part during CompleteMultipartUpload. (#1518) (#1538) 2016-05-08 23:49:49 -07:00
Harshavardhana 90ea494338 erasure: waitCloser should implement CloseWithError. (#1543) 
This is needed so that the other end of the pipe receives
and error, cleanups temporary files.
 2016-05-08 16:26:10 -07:00
Harshavardhana a8fdd04e62 erasure: ReadFile should honor proper offsets. (#1542) 
Fixes #1535
 2016-05-08 15:39:24 -07:00
Harshavardhana 76c511c9fe api: Extend S3 errors with Minio errors. (#1533) 
Fixes #1530
 2016-05-08 12:36:16 -07:00
Krishna Srinivas 75320f70d0 multipart: reject part upload if size is less than 5MB. (#1518) 2016-05-08 12:06:05 -07:00
Krishna Srinivas 88e1c04259 XL/ListDir: break out of loop if list on one disk is a success. (#1534) 2016-05-08 12:05:19 -07:00
Krishna Srinivas a205aca6d2 init: Cleanup .minio/tmp directories recursively. Also takes care of cleaning up of parts directory during abortMultipartUpload. (#1532) 2016-05-08 10:15:34 -07:00
Harshavardhana 3f51dd4fd4 xl: CompleteMultipartUpload should rename files in a routine. (#1527) 
This solves the client timeout while renaming 9000+ parts.

Fixes #1526
 2016-05-08 02:38:35 -07:00
Harshavardhana 56b7df90e1 xl/fs: ListObjectParts should set nextPartNumberMarker properly. (#1528) 
For list requests on parts more than 1000, would lead to an infinite
loop.

Fixes #1522
 2016-05-08 02:21:12 -07:00
Harshavardhana a56d5ef415 xl/fs: isFunctions should only return boolean. (#1525) 
log the unrecognize errors.
 2016-05-08 01:58:05 -07:00
Harshavardhana 937d68202d server: Deadcode removal. (#1517) 2016-05-07 21:47:33 -07:00
Harshavardhana bf563afb80 xl: DeleteObject regression over errChannel. (#1521) 
DeleteObject would hang indefinitely - fix it.
 2016-05-07 12:48:12 -07:00
Harshavardhana 091c1e8456 copyObject: No need to verify md5sum. (#1520) 
Multipart objects are kept in non hex md5sum format.
This format doesn't comply with hex, so decoding
would fail invariably.

This is not necessary to validate and its not expected
error during a CopyObject operation.

Fixes #1519
 2016-05-07 03:43:08 -07:00
Harshavardhana 751fa972f5 xl/fs: Multipart re-org introduce "uploads.json" (#1505) 
Fixes #1457
 2016-05-07 02:08:03 -07:00
Harshavardhana 434423de89 xl: Move format detection inside xl objects. (#1515) 
Fixes #1449
 2016-05-07 00:59:43 -07:00
Harshavardhana a20ccb1e83 server: Print proper endpoint, along with https if configured. (#1514) 
Fixes #1492
 2016-05-06 21:18:29 -07:00
Harshavardhana 0625c050e6 xl/tests: Enable server handler tests over XL. (#1512) 
Fixes #1513
 2016-05-06 16:47:23 -07:00
Harshavardhana 0b74f5624e xl: Fix how we deal with read offsets at erasure layer. (#1511) 
Requires skipping necessary parts of dataBlocks during
decoding phase and requires us to properly skip the
entries as needed.

Thanks to Karthic for reproducing this important issue.

Fixes #1503
 2016-05-06 16:25:08 -07:00
Krishna Srinivas c06b9abc15 bucket-handlers: do not unescape marker as gorilla layer would have already done it. (#1495) (#1510) 2016-05-06 16:04:46 -07:00
Krishna Srinivas a5d31d4254 XL/ListObjects: use string.TrimSuffix instead of Trim. (#1498) (#1509) 2016-05-06 13:28:55 -07:00
karthic rao 20ca65c793 Cleanup: mispell fixes 2016-05-06 12:32:44 -07:00
karthic rao 0b4bbe6d9e Adding XL Object layer validation for existing unit tests of single node (#1499) 
object layer.

Adding isBucketExist check for GetObjectInfo in the XL layer.
 2016-05-06 11:57:04 -07:00
Krishna Srinivas 48d3be36da XL/ListObjects: Fix ordering issue during listing if the files were uploaded as multipart uploads. (#1498) (#1506) 
i.e if two files "tmp" and "tmp.1" are uploaded as multipart we would list ""tmp.1"" before ""tmp"" as "tmp.1/" < "tmp/"
 2016-05-06 10:19:09 -07:00
Harshavardhana 5133ea50bd xl/fs: Make i/o operations atomic. (#1496) 2016-05-05 20:28:22 -07:00
Harshavardhana 17868ccd7f handlers: overhaul entire writErrorResponse, simplify. (#1472) 2016-05-05 20:24:29 -07:00
Harshavardhana ba5805e60a bucketPolicy: Do not use regexes, just do prefix matches. (#1497) 
AWS arn supports wildcards and this is flat namespace, simple
prefix matching is fine.

Fixes #1481
Fixes #1482
 2016-05-05 19:58:48 -07:00
Harshavardhana ca097de96c xl/fs: Add initObjectLayer function. (#1494) 
Fixes #1493
 2016-05-05 15:00:03 -07:00
Bala FA 658a595d7a xl-erasure: RenameFile should support quorum. (#1487) 
Fixes #1463
 2016-05-05 13:03:53 -07:00
Krishna Srinivas 247e835d7b object: move go-routine listing from posix to objectLayer. (#1491) 2016-05-05 12:51:56 -07:00
Harshavardhana 46680788f9 xl/fs: cleanup '/.minio/tmp' directory on each initialization. (#1490) 2016-05-05 01:54:43 -07:00
Harshavardhana ad40036cba posix: filepath shouldn't be used anymore use path.Join (#1486) 2016-05-05 01:39:26 -07:00
Harshavardhana 82fbe908a3 object: DeleteBucket should return proper error for BucketNotEmpty. (#1489) 
Fixes #1488
 2016-05-05 01:35:21 -07:00
Harshavardhana f145e1042f quick/config: No need to use Data() with type assertion. (#1480) 
Since input to quick.New() is a pointer the unmarshalled value
internally already has the value, subsequent type assertions
are not needed.

Thanks to Bala for finding this behavior.

Fixes #1475
 2016-05-04 20:22:15 -07:00