mc rm to support entire namespace removal #2265

poornas · 2017-09-29T21:04:17Z

Fixes #2250
Currently, mc rm --recursive --force works on buckets but not on the entire site. This PR adds support for rm --recursive to delete entire site.

To distinguish full site removal from removing a bucket and provide appropriate warning to the user, the user needs to supply --dangerous flag in addition to the --recursive and --force flags for full site rm.

The rm --recursive --force option will continue to work as before at the bucket level or on the file system.

codecov-io · 2017-09-29T21:09:09Z

Codecov Report

Merging #2265 into master will decrease coverage by 0.18%.
The diff coverage is 0%.

@@            Coverage Diff             @@
##           master    #2265      +/-   ##
==========================================
- Coverage   10.95%   10.76%   -0.19%     
==========================================
  Files         102      102              
  Lines        7624    10042    +2418     
==========================================
+ Hits          835     1081     +246     
- Misses       6654     8823    +2169     
- Partials      135      138       +3

Impacted Files	Coverage Δ
cmd/rm-main.go	`0% <0%> (ø)`	⬆️
cmd/client-s3.go	`13.89% <0%> (-0.68%)`	⬇️
cmd/pipechan.go	`87.8% <0%> (-12.2%)`	⬇️
cmd/session.go	`35.89% <0%> (-6.48%)`	⬇️
pkg/hookreader/hookreader.go	`19.23% <0%> (-5.77%)`	⬇️
cmd/humanized-duration.go	`59.57% <0%> (-4.53%)`	⬇️
cmd/client-url.go	`57.85% <0%> (-2.33%)`	⬇️
pkg/probe/probe.go	`42.59% <0%> (-2.24%)`	⬇️
cmd/ls.go	`10.71% <0%> (-1.45%)`	⬇️
cmd/config.go	`20.56% <0%> (-1.4%)`	⬇️
... and 90 more

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 65c582c...c65be4d. Read the comment docs.

harshavardhana · 2017-09-29T21:34:28Z

cmd/rm-main.go

+      $ {{.HelpName}} --recursive --namespace s3
+
+   6. Remove all buckets and objects older than '90' days from entire namespace
+			$ {{.HelpName}} --recursive --namespace --older-than=90 s3


Tabbing is off.

balamurugana · 2017-10-03T03:23:59Z

cmd/rm-main.go

+	isNamespace := ctx.Bool("namespace")
+	isNamespaceRemoval := false
+	for _, url := range ctx.Args() {
+		if isHostName(url) {


This method does not work for removing a directory with non-aliases. For example mc rm mylocaldir would fail. Basically we need to differentiate alias and local dir

changed this to incorporate feedback

balamurugana · 2017-10-03T03:27:46Z

cmd/rm-main.go

-
+	if (isRecursive || isStdin) && isNamespaceRemoval && !isNamespace {
+		fatalIf(errDummy().Trace(),
+			"Full namespace removal requires --namespace option. This operation removes all the buckets on your namespace. Please review carefully before performing this *DANGEROUS* operation.")


This check is not interesting. removing local file system, we don't need any restriction like that.

balamurugana · 2017-10-03T03:31:29Z

cmd/rm-main.go

+		// print the rmMessage only if content is a bucket or object. clnt.List lists
+		// prefixes as well which unifies the S3 listing with folder listing - however
+		// these prefixes are not actual objects on s3 and the content.Time defaults to
+		// zero. Filter these out to prevent erroneous rmMessage being printed to console.


You can skip entries those are ending with / and having zero time for completeness.

balamurugana · 2017-10-03T03:38:16Z

cmd/rm-main.go

+					if !ok {
+						close(contentCh)
+						return exitStatus(globalErrorExitStatus)
+					}


We would need to exit on first error here. This check leads to continuation. Below is the right logic

case pErr, ok := <-errorCh: if ok { errorIf(pErr.Trace(urlString), "Failed to remove `"+urlString+"`.") switch pErr.ToGoError().(type) { case PathInsufficientPermission: // Ignore Permission error. continue } } close(contentCh) return exitStatus(globalErrorExitStatus)

IMO, having ok is not required because pErr will be nil on closed channel and we exit anyway.

fixed this.

balamurugana · 2017-10-03T03:46:28Z

cmd/rm-main.go

+func removeSiteRecursive(url string, isIncomplete bool, isFake bool, older int) (err error) {
+	clnt, pErr := newClient(url)
+	if pErr != nil {
+		errorIf(pErr.Trace(url), "Failed to remove `"+url+"` recursively.")


This is not appropriate error message here.

This is fine because of pErr.Trace(url). Sorry for the confusion.

balamurugana · 2017-10-03T03:57:25Z

cmd/rm-main.go

-			err = removeRecursive(url, isIncomplete, isFake, older)
+			if isHostName(url) {
+				if isNamespace {
+					err = removeSiteRecursive(url, isIncomplete, isFake, older)


Unable to understand the need of this function.

Current code fails for below commands

$ mc rm --recursive --fake --force play mc: <ERROR> Failed to remove `play` recursively. Bucket `` does not exist.

What I am able to visualize with --namespace flag is by running mc rm --recursive --fake --force play, it could get list of buckets from play alias and keep calling removeRecursive() with each bucket URL.

harshavardhana · 2017-10-05T04:53:16Z

cmd/rm-main.go

@@ -264,6 +293,43 @@ func removeRecursive(url string, isIncomplete bool, isFake bool, older int) erro
 	return nil
 }

+func isHostName(url string) bool {


Add unit tests?

removed this function.

balamurugana · 2017-10-05T23:47:52Z

cmd/rm-main.go

@@ -245,7 +246,7 @@ func removeRecursive(url string, isIncomplete bool, isFake bool, older int) erro
 		// prefixes as well which unifies the S3 listing with folder listing - however
 		// these prefixes are not actual objects on s3 and the content.Time defaults to
 		// zero. Filter these out to prevent erroneous rmMessage being printed to console.
-		if !content.Time.IsZero() {
+		if !(content.Time.IsZero() && strings.HasSuffix(urlString, string(filepath.Separator))) {


S3 service sends / as separator whereas on windows \. We would need to check whether this works on windows or not.

balamurugana · 2017-10-05T23:48:40Z

cmd/rm-main.go

@@ -351,7 +344,7 @@ func mainRm(ctx *cli.Context) error {
 	// Support multiple targets.
 	for _, url := range ctx.Args() {
 		if isRecursive {
-			if isHostName(url) {
+			if !isAliasURLDir(url) && (strings.Count(url, string(filepath.Separator)) == 0) {


same as above

isAliasURLDir is used internally by the checkCopySyntax function. Here the intent is to make sure only the alias is entered, and it is at the root level not a subdirectory.

harshavardhana

You should look at the cp-url-syntax.go implementation to understand the how to guess the URL types.

checkCopySyntax

poornas · 2017-10-25T23:04:18Z

@harshavardhana and @balamurugana , updated the PR with a different approach - recursive removal is handled agnostic to fs/s3.

harshavardhana

There is some inconsistent error messaging here

$ mc rm --recursive s3
mc: <ERROR> Removal requires --force option. This operation is *IRREVERSIBLE*. Please review carefully before performing this *DANGEROUS* operation.

$ mc rm --recursive --force s3
mc: <ERROR> Full site removal requires --dangerous option. This operation removes all the buckets on your namespace. Please review carefully before performing this *DESTRUCTIVE* operation.

Requires like two steps to get to --dangerous

$ mc rm --recursive --force --dangerous s3
Removing `s3/aead/manual.txt`.
Removing `s3/aead/my-encrypted-object.txt`.
Removing `s3/aead`.
Removing `s3/ersan-ersan`.
Removing `s3/kannappan/list.go`.
Removing `s3/kannappan/1/2/3`.
Removing `s3/kannappan/a/b/`.
Removing `s3/kannappan`.
Removing `s3/kannappan1`.
Removing `s3/mc-mint-test-bucket-10442/2123123\123`.
Removing `s3/mc-mint-test-bucket-10442`.
panic: close of nil channel

goroutine 7 [running]:
github.com/minio/mc/cmd.(*s3Client).Remove.func1(0xc4200860c0, 0xc420086060, 0xc420349480, 0xc42030fd00, 0xc42000ea30, 0xc42000ea38, 0x405000, 0xc42000ea40, 0xc42005c15c)
	/home/harsha/mygo/src/github.com/minio/mc/cmd/client-s3.go:723 +0x3ff
created by github.com/minio/mc/cmd.(*s3Client).Remove
	/home/harsha/mygo/src/github.com/minio/mc/cmd/client-s3.go:666 +0x144

Also there seems to be a bug when doing the entire removal, looks like due to an error.

harshavardhana · 2017-11-01T08:06:18Z

To distinguish full site removal from removing a bucket and provide appropriate warning to the user, the user needs to supply --dangerous flag in addition to the --recursive and --force flags for full site rm.

Tested the PR mc rm -r --force play/ seems to delete everything, but strangely --dangerous flag is removed was that intended?

poornas · 2017-11-01T16:49:58Z

that was an intermediate commit - updated the PR to fix messaging as well.

harshavardhana · 2017-11-05T07:15:36Z

Can you sqaush the commits here,. Will test once again after they are squashed.

poornas · 2017-11-05T07:51:36Z

done.

harshavardhana

LGTM, minor changes.

harshavardhana · 2017-11-05T18:23:23Z

cmd/rm-main.go

@@ -82,16 +88,22 @@ EXAMPLES:
   1. Remove a file.
      $ {{.HelpName}} 1999/old-backup.tgz

-   2. Remove all objects recursively.
+   2. Remove all objects recursively from a bucket.


Remove all objects recursively from bucket 'jazz-songs' matching 'louis' prefix.

harshavardhana · 2017-11-05T18:23:39Z

cmd/rm-main.go

      $ {{.HelpName}} --recursive s3/jazz-songs/louis/

-   3. Remove all objects older than '90' days.
+   3. Remove all objects older than '90' days from a bucket.


Same as above but with 'older than 90 days'

harshavardhana · 2017-11-05T18:25:47Z

cmd/rm-main.go

      $ {{.HelpName}} --recursive --older-than=90 s3/jazz-songs/louis/

   4. Remove all objects read from STDIN.
      $ {{.HelpName}} --force --stdin

-   5. Drop all incomplete uploads on 'jazz-songs' bucket.
+   5. Remove all buckets and objects recursively from host


Remove all buckets and objects recursively from S3 host

poornas · 2017-11-13T23:49:33Z

@harshavardhana and @balamurugana , review comments were addressed. needs a revisit.

harshavardhana

LGTM tested

balamurugana

minor comments

balamurugana · 2017-11-15T04:55:08Z

cmd/rm-main.go

+	isDangerous := ctx.Bool("dangerous")
+	isNamespaceRemoval := false
+	for _, url := range ctx.Args() {
+		if !isAliasURLDir(url) && (strings.Count(url, string(filepath.Separator)) == 0) {


Please add a comment about meaning of this logic.

() around strings. is not required.

balamurugana · 2017-11-15T04:58:07Z

cmd/rm-main.go

-			Size: content.Size,
-		})
-
+		if !(content.Time.IsZero() && strings.HasSuffix(urlString, string(filepath.Separator))) {


Please add a comment about this check

balamurugana · 2017-11-15T22:23:25Z

cmd/rm-main.go

+	for _, url := range ctx.Args() {
+		// check if url represents just the namespace without bucket or object prefix specified
+		// in the url
+		if !isAliasURLDir(url) && strings.Count(url, string(filepath.Separator)) == 0 {


A valid input s3/ and unusual input s3// wouldn't work here. We need to take care properly i.e. we shouldn't use strings.Count() inside use url to alias conversion.

If mentioned input is not supported, we have to error out properly.

you are correct @balamurugana , added additional validation so that user input errors are caught right away.

balamurugana · 2017-11-16T23:00:03Z

cmd/rm-main.go

+		}
+		if alias != "" && path == "" && isBucketEmpty {
+			isNamespaceRemoval = true
+			break


The whole checker logic is confusing to me. In rm we should not check for bucket or object which is completely abstracted by client interface. I see below logic is good enough

url = filepath.ToSlash(url) if !isAliasURLDir(url) { alias, path := url2Alias(url) isNamespaceRemoval = (path == "") }

Also you need to aware that ctx.Args() may come as multiple different values for single run e.g.
ctx.Args() = ["mysite", "play/my-bucket", "play/another-bucket/my/prefix", "mydir", "mydir2/my/path"]

without the check on lines 157-158, errors like mc rm play// wont be caught. Earlier it was being caught in s3 api because bucket name was not permitted to be empty. Now by allowing namespace removal, bucket name can be empty and there is no way to validate cases like multiple consecutive / delimiters being entered as input. Lines 160-163 can be replaced by the logic you proposed.

I see url = filepath.ToSlash(url) is wrong, you should use url = filepath.Clean(url)

balamurugana · 2017-11-18T00:05:09Z

cmd/rm-main.go

+	isDangerous := ctx.Bool("dangerous")
+	isNamespaceRemoval := false
+	for _, url := range ctx.Args() {
+		url = filepath.Clean(url)


you could use path.Clean() to work with MS Windows.

doesn't filepath.Clean do the same?

path.Clean() doesn't do anything on windows @balamurugana @poornas - filepath.Clean() should be used for cross platform friendly behavior.

See below. This had been discussed long back

package main import ( "fmt" "path" "path/filepath" ) func main() { fmt.Println("filepath.Clean():") fmt.Println(filepath.Clean("/s3")) fmt.Println(filepath.Clean("s3/")) fmt.Println(filepath.Clean("s3/mybucket//object//")) fmt.Println(filepath.Clean(`\s3`)) fmt.Println(filepath.Clean(`s3\`)) fmt.Println(filepath.Clean(`s3\mybucket\\object\\`)) fmt.Println("path.Clean():") fmt.Println(path.Clean("/s3")) fmt.Println(path.Clean("s3/")) fmt.Println(path.Clean("s3/mybucket//object//")) fmt.Println(path.Clean(`\s3`)) fmt.Println(path.Clean(`s3\`)) fmt.Println(path.Clean(`s3\mybucket\\object\\`)) }

On MS Windows

filepath.Clean(): \s3 s3 s3\mybucket\object \s3 s3 s3\mybucket\object path.Clean(): /s3 s3 s3/mybucket/object \s3 s3\ s3\mybucket\\object\\

On GNU/Linux

filepath.Clean(): /s3 s3 s3/mybucket/object \s3 s3\ s3\mybucket\\object\\ path.Clean(): /s3 s3 s3/mybucket/object \s3 s3\ s3\mybucket\\object\\

i don't think so you understand the problem of using path.Clean() , path.Clean() doesn't work on windows since it never cleans paths such as \\ or C:\mypath\...\otherdir\ . path.Clean() doesn't understand C:\ etc. so it would be wrong to use path.Clean() for cross platform reasons.

For mc we do support such paths and need to behave properly for windows natively once we do filepath.Clean() if we want all the \ to / then we should use filepath.ToSlash() to convert the paths properly.

Using code like in this manner filepath.ToSlash(filepath.Clean(p)) would work for all operating systems.

How is it possible to store object name containing \ with filepath.ToSlash(filepath.Clean(p))?

If this is unsupported by mc, go ahead and do the change you mentioned. Also check whether you handle UNC path correctly.

filepath.ToSlash is used to do proper Windows translation, filepath.ToSlash has no effect on Unixes. Now if you are arguing that if there are files which have \ on Linux then they are stored fine with this mechanism as well. Now on windows there is no filename which explicitly contains \ since its a separator.

Here is how the behavior is

filepath.ToSlash converts all \ on Windows to /

filepath.ToSlash converts all / to / on Unixes.

filepath.ToSlash is not ubiquitously converting \ to / it is a platform dependent behavior - https://play.golang.org/p/fFPdEZoAEf

balamurugana

LGTM

harshavardhana · 2017-11-28T08:23:52Z

@poornas can you squash your changes? would it possible to bring in some unit tests as well if possible?

harshavardhana · 2017-11-28T08:27:32Z

cmd/rm-main.go

+	isDangerous := ctx.Bool("dangerous")
+	isNamespaceRemoval := false
+	for _, url := range ctx.Args() {
+		url = path.Clean(url)


It is required to perhaps bring change like this instead filepath.ToSlash(filepath.Clean(url)) here, for windows paths such as C:\ path.Clean will not work.

Also question for me why is this path.Clean needed? is it for isAliasURLDir() ? A comment would be needed i think here.

@harshavardhana , on further thinking, the path.Clean and isAliasURLDir() check are superfluous and can be dispensed with altogether.

Okay that would be better.

balamurugana

Revoking +1 till finalize on path.Clean() or filepath.Clean() on windows. Also test with UNC path as well.

poornas · 2017-11-28T19:09:24Z

@harshavardhana , have not introduced stand alone functions that can be unit tested here. It might be worthwhile to add a mint cleanup function in the functional tests that exercises this code - but that will cleanup play server and may not be entirely good.

harshavardhana · 2017-11-28T19:38:11Z

@harshavardhana , have not introduced stand alone functions that can be unit tested here. It might be worthwhile to add a mint cleanup function in the functional tests that exercises this code - but that will cleanup play server and may not be entirely good.

Okay that makes sense.

harshavardhana

LGTM

balamurugana · 2017-11-28T21:46:22Z

cmd/rm-main.go

+	isDangerous := ctx.Bool("dangerous")
+	isNamespaceRemoval := false
+	for _, url := range ctx.Args() {
+		_, path := url2Alias(url)


What are all expected behaviors for below command line arguments?

1. mc rm mylocaldir 2. mc rm mylocaldir/ 3. mc rm mylocalfile 4. mc rm c: 5. mc rm c:\ 6. mc rm \\myserver\myfolder\myfile

➜ mc git:(rr2) mc rm mylocaldir Removing `mylocaldir`. ➜ mc git:(rr2) mkdir mylocaldir ➜ mc git:(rr2) touch mylocaldir/testfile ➜ mc git:(rr2) ✗ mc rm mylocaldir Removing `mylocaldir`. mc: <ERROR> Failed to remove `mylocaldir`. remove mylocaldir: directory not empty ➜ mc git:(rr2) ✗ mc rm mylocaldir Removing `mylocaldir`. mc: <ERROR> Failed to remove `mylocaldir`. remove mylocaldir: directory not empty

however, removing the isAliasURLDir() check will interfere with mc rm messaging in recursive removal. Which is why the check was introduced in the first place - for filtering out fs.

➜ mc git:(rr2) ✗ mc rm --r --force mylocaldir mc: <ERROR> Full site removal requires --dangerous option. This operation removes all the buckets on your namespace. Please review carefully before performing this *DESTRUCTIVE* operation. ```

@harshavardhana and @balamurugana , added the isAliasURLDir() check back and a comment on why it is needed.

balamurugana · 2017-11-28T22:46:55Z

cmd/rm-main.go

+	for _, url := range ctx.Args() {
+		// namespace removal applies only for non FS. So filter out if passed url represents a directory
+		if !isAliasURLDir(url) {
+			_, path := url2Alias(url)


Does url2Alias() take care of returning clean path including windows UNC path?

yes, see test output below..

C:\go\bin> cd ..\src\github.com\minio\mc C:\go\src\github.com\minio\mc [rr2 ≡]> mc rm --r --force \\myserver\myfolder\myfile mc.exe: <ERROR> Failed to remove `\\myserver\myfolder\myfile` recursively. open \\myserver\myfolder\myfile: The network name cannot be found. C:\go\src\github.com\minio\mc [rr2 ≡]> mc rm --r --force C: Removing `C:.git\FETCH_HEAD`. C:\go\src\github.com\minio\mc> mc rm C:\ Removing `C:\`. mc.exe: <ERROR> Failed to remove `C:\`. Insufficient permissions to access this file `C:\` C:\go\src\github.com\minio\mc> C:\go\src\github.com\minio\mc> mc rm C: Removing `C:`. mc.exe: <ERROR> Failed to remove `C:`. remove C:: The process cannot access the file because it is being used by a ther process. C:\go\src\github.com\minio\mc> C:\go\src\github.com\minio\mc> mc rm --r --force C: Removing `C:cmd\mc_test.go`. ~> mc mb a/b/c/d ~> mc rm a Removing `a`. mc.exe: <ERROR> Failed to remove `a`. remove a: The directory is not empty. ~> mc rm \\TMP\ss\ mc.exe: <ERROR> Failed to remove `\\TMP\ss\`. GetFileAttributesEx \\TMP\ss: The network name cannot be ~> mc rm C:\Users\vagrant\a\\\ Removing `C:\Users\vagrant\a\\\`. mc.exe: <ERROR> Failed to remove `C:\Users\vagrant\a\\\`. remove C:\Users\vagrant\a\\\: The directory ~> mc rm C:\Users\vagrant\a\\\b\\\c\\\d Removing `C:\Users\vagrant\a\\\b\\\c\\\d`. ~>

We used path.Clean() to take care aliases like s3 and s3/. Hope return path from url2Alias() is clean.

yes, added a comment for that as well.

balamurugana · 2017-11-29T09:48:59Z

cmd/rm-main.go

+	isNamespaceRemoval := false
+	for _, url := range ctx.Args() {
+		// clean path for aliases like s3/
+		url = filepath.ToSlash(filepath.Clean(url))


interestingly UNC path using / is working in go1.9.2 by breaking the UNC path specification. I would recommend to add this comment as you convert any path separator to / here.

It is not breaking @balamurugana they automatically convert it internally properly. Before any Win32 calls are made os package converts them to necessary values.

FWIW Win32 calls support / path separators as well.

Check on windows how it behaves UNC path with /

C:\GO\src\github.com\minio\mc [(0f4e134...) +0 ~1 -0 !]> mc mb a/b/c/d/e Bucket created successfully `a/b/c/d/e`. C:\GO\src\github.com\minio\mc [(0f4e134...) +0 ~1 -0 !]> tree a Folder PATH listing for volume Windows 10 Volume serial number is 00000200 4286:2744 C:\GO\SRC\GITHUB.COM\MINIO\MC\A └───b └───c └───d └───e C:\GO\src\github.com\minio\mc [(0f4e134...) +0 ~1 -0 !]> mc rm a/b/c/d/e Removing `a/b/c/d/e`. C:\GO\src\github.com\minio\mc [(0f4e134...) +0 ~1 -0 !]> C:\GO\src\github.com\minio\mc [(0f4e134...) +0 ~1 -0 !]> mc rm //network1/blah/blah mc.exe: <ERROR> Failed to remove `//network1/blah/blah`. Requested file `/blah/blah` not found C:\GO\src\github.com\minio\mc [(0f4e134...) +0 ~1 -0 !]> mc rm --r --force //network1/blah/blah mc.exe: <ERROR> Failed to remove `//network1/blah/blah` recursively. open /blah/blah: The system canno h specified.

No surprise go applications are working cleanly. See my above comment. Just add a comment what exactly i mentioned.

harshavardhana · 2017-11-30T02:08:38Z

Moving this to blocked as requested by @poornas until @abperiasamy gives final review on the changes and UI.

balamurugana

LGTM

This requires --dangerous flag in addition to --recursive and --force Change namespace identification check to be more robust to user input errors like mc rm --r --force play//

poornas · 2017-12-02T02:18:05Z

after discussion with @abperiasamy , changed messaging for the --force and --dangerous flags and error message if flags are not supplied for site wide rm command. This is unblocked now

harshavardhana reviewed Sep 29, 2017

View reviewed changes

poornas force-pushed the rr2 branch 3 times, most recently from 11fd120 to 1b66a05 Compare September 29, 2017 21:42

deekoder requested review from harshavardhana and balamurugana October 2, 2017 20:16

balamurugana requested changes Oct 3, 2017

View reviewed changes

harshavardhana requested changes Oct 5, 2017

View reviewed changes

balamurugana reviewed Oct 5, 2017

View reviewed changes

harshavardhana reviewed Oct 7, 2017

View reviewed changes

poornas force-pushed the rr2 branch 2 times, most recently from b2a0a38 to 42f05f9 Compare October 25, 2017 22:55

harshavardhana mentioned this pull request Oct 31, 2017

mc rm not erroring out when deleting directory #2242

Closed

harshavardhana requested changes Oct 31, 2017

View reviewed changes

poornas force-pushed the rr2 branch from 42f05f9 to 9fe4d50 Compare October 31, 2017 23:24

poornas force-pushed the rr2 branch from 2512cbe to b84dd9e Compare November 2, 2017 20:01

poornas force-pushed the rr2 branch from b84dd9e to ae64d13 Compare November 5, 2017 07:34

harshavardhana requested changes Nov 5, 2017

View reviewed changes

poornas force-pushed the rr2 branch 2 times, most recently from a79ab47 to 0c847db Compare November 5, 2017 18:35

harshavardhana approved these changes Nov 14, 2017

View reviewed changes

balamurugana reviewed Nov 15, 2017

View reviewed changes

poornas force-pushed the rr2 branch from 0c847db to 84dbade Compare November 15, 2017 18:04

balamurugana reviewed Nov 15, 2017

View reviewed changes

balamurugana reviewed Nov 16, 2017

View reviewed changes

poornas force-pushed the rr2 branch from eea4fbe to 455643e Compare November 17, 2017 03:06

balamurugana reviewed Nov 18, 2017

View reviewed changes

poornas force-pushed the rr2 branch from 455643e to 75a00ed Compare November 18, 2017 00:50

balamurugana approved these changes Nov 28, 2017

View reviewed changes

harshavardhana requested changes Nov 28, 2017

View reviewed changes

balamurugana requested changes Nov 28, 2017

View reviewed changes

poornas force-pushed the rr2 branch from 75a00ed to 3479f7e Compare November 28, 2017 18:53

harshavardhana approved these changes Nov 28, 2017

View reviewed changes

balamurugana reviewed Nov 28, 2017

View reviewed changes

poornas force-pushed the rr2 branch from 3479f7e to 0f4e134 Compare November 28, 2017 22:40

balamurugana reviewed Nov 28, 2017

View reviewed changes

poornas force-pushed the rr2 branch from 0f4e134 to 1e4cad4 Compare November 29, 2017 01:19

harshavardhana approved these changes Nov 29, 2017

View reviewed changes

balamurugana reviewed Nov 29, 2017

View reviewed changes

poornas force-pushed the rr2 branch from 1e4cad4 to cec5e4e Compare November 30, 2017 02:05

harshavardhana added the pending discussion label Nov 30, 2017

balamurugana approved these changes Nov 30, 2017

View reviewed changes

Add mc rm support to recursively remove site

c65be4d

This requires --dangerous flag in addition to --recursive and --force Change namespace identification check to be more robust to user input errors like mc rm --r --force play//

poornas force-pushed the rr2 branch from cec5e4e to c65be4d Compare December 2, 2017 02:10

harshavardhana removed the pending discussion label Dec 2, 2017

harshavardhana approved these changes Dec 2, 2017

View reviewed changes

harshavardhana merged commit 1e7b4c9 into minio:master Dec 2, 2017

mc rm to support entire namespace removal #2265

mc rm to support entire namespace removal #2265

Conversation

poornas commented Sep 29, 2017 • edited Loading

codecov-io commented Sep 29, 2017 • edited Loading

Codecov Report

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

harshavardhana left a comment

Choose a reason for hiding this comment

poornas commented Oct 25, 2017

harshavardhana left a comment

Choose a reason for hiding this comment

harshavardhana commented Nov 1, 2017

poornas commented Nov 1, 2017

harshavardhana commented Nov 5, 2017

poornas commented Nov 5, 2017

harshavardhana left a comment • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

poornas commented Nov 13, 2017

harshavardhana left a comment

Choose a reason for hiding this comment

balamurugana left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

balamurugana Nov 16, 2017 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

balamurugana Nov 19, 2017 • edited Loading

Choose a reason for hiding this comment

harshavardhana Nov 19, 2017 • edited Loading

Choose a reason for hiding this comment

balamurugana left a comment

Choose a reason for hiding this comment

harshavardhana commented Nov 28, 2017

Choose a reason for hiding this comment

poornas Nov 28, 2017 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

balamurugana left a comment

Choose a reason for hiding this comment

poornas commented Nov 28, 2017

harshavardhana commented Nov 28, 2017

harshavardhana left a comment

Choose a reason for hiding this comment

balamurugana Nov 28, 2017 • edited Loading

Choose a reason for hiding this comment

poornas Nov 28, 2017 • edited Loading

Choose a reason for hiding this comment

poornas commented Sep 29, 2017 •

edited

Loading

codecov-io commented Sep 29, 2017 •

edited

Loading

harshavardhana left a comment •

edited

Loading

balamurugana Nov 16, 2017 •

edited

Loading

balamurugana Nov 19, 2017 •

edited

Loading

harshavardhana Nov 19, 2017 •

edited

Loading

poornas Nov 28, 2017 •

edited

Loading

balamurugana Nov 28, 2017 •

edited

Loading

poornas Nov 28, 2017 •

edited

Loading

poornas Nov 28, 2017 •

edited

Loading

balamurugana Nov 29, 2017 •

edited

Loading