Merge branch '18363-managed-properties-doc-improvement' into main.
authorLucas Di Pentima <lucas.dipentima@curii.com>
Wed, 17 Nov 2021 18:28:37 +0000 (15:28 -0300)
committerLucas Di Pentima <lucas.dipentima@curii.com>
Wed, 17 Nov 2021 18:28:37 +0000 (15:28 -0300)
Closes #18363

Arvados-DCO-1.1-Signed-off-by: Lucas Di Pentima <lucas.dipentima@curii.com>

12 files changed:
AUTHORS
CONTRIBUTING.md
doc/user/topics/collection-versioning.html.textile.liquid
lib/config/export.go
sdk/cwl/setup.py
sdk/python/arvados/collection.py
sdk/python/setup.py
sdk/python/tests/run_test_server.py
sdk/python/tests/test_collections.py
services/api/app/controllers/arvados/v1/groups_controller.rb
services/api/test/functional/arvados/v1/groups_controller_test.rb
services/keepstore/unix_volume.go

diff --git a/AUTHORS b/AUTHORS
index 93bdb37dc8f89fa0e8452804beee4943faa9c22e..b8b75518ff08ad0f0d787f27f52ba8f58f2bece2 100644 (file)
--- a/AUTHORS
+++ b/AUTHORS
@@ -20,4 +20,5 @@ Chen Chen <aflyhorse@gmail.com>
 Veritas Genetics, Inc. <*@veritasgenetics.com>
 Curii Corporation, Inc. <*@curii.com>
 Dante Tsang <dante@dantetsang.com>
-Codex Genetics Ltd <info@codexgenetics.com>
\ No newline at end of file
+Codex Genetics Ltd <info@codexgenetics.com>
+Bruno P. Kinoshita <brunodepaulak@yahoo.com.br>
index 39483ce62d879d5e7c8ba645315b0041f5271bd1..3376311e5c80690cd6088e951d6d9582e504e2d0 100644 (file)
@@ -14,19 +14,19 @@ hear about Arvados success stories.
 Those interested in contributing should begin by joining the [Arvados community
 channel](https://gitter.im/arvados/community) and telling us about your interest.
 
-Contributers should also create an account at https://dev.arvados.org
+Contributors should also create an account at https://dev.arvados.org
 to be able to create and comment on bug tracker issues.  The
 Arvados public bug tracker is located at
 https://dev.arvados.org/projects/arvados/issues .
 
-Contributers may also be interested in the [development road map](https://dev.arvados.org/issues/gantt?utf8=%E2%9C%93&set_filter=1&gantt=1&f%5B%5D=project_id&op%5Bproject_id%5D=%3D&v%5Bproject_id%5D%5B%5D=49&f%5B%5D=&zoom=1).
+Contributors may also be interested in the [development road map](https://dev.arvados.org/issues/gantt?utf8=%E2%9C%93&set_filter=1&gantt=1&f%5B%5D=project_id&op%5Bproject_id%5D=%3D&v%5Bproject_id%5D%5B%5D=49&f%5B%5D=&zoom=1).
 
 # Development
 
 Git repositories for primary development are located at
 https://git.arvados.org/ and can also be browsed at
 https://dev.arvados.org/projects/arvados/repository .  Every push to
-the master branch is also mirrored to Github at
+the main branch is also mirrored to Github at
 https://github.com/arvados/arvados .
 
 Visit [Hacking Arvados](https://dev.arvados.org/projects/arvados/wiki/Hacking) for
@@ -46,12 +46,12 @@ This is the general contribution process:
 2. Clone your fork, make your changes, commit to your fork.
 3. Every commit message must have a DCO sign-off and every file must have a SPDX license (see below).
 4. Add yourself to the [AUTHORS](AUTHORS) file
-5. When your fork is ready, through Github, Create a Pull Request against `arvados:master`
+5. When your fork is ready, through Github, Create a Pull Request against `arvados:main`
 6. Notify the core team about your pull request through the [Arvados development
 channel](https://gitter.im/arvados/development) or by other means.
 7. A member of the core team will review the pull request.  They may have questions or comments, or request changes.
 8. When the contribution is ready, a member of the core team will
-merge the pull request into the master branch, which will
+merge the pull request into the main branch, which will
 automatically resolve the pull request.
 
 The Arvados project does not require a contributor agreement in advance, but does require each commit message include a [Developer Certificate of Origin](https://dev.arvados.org/projects/arvados/wiki/Developer_Certificate_Of_Origin).  Please ensure *every git commit message* includes `Arvados-DCO-1.1-Signed-off-by`. If you have already made commits without it, fix them with `git commit --amend` or `git rebase`.
@@ -68,7 +68,7 @@ New files must also include `SPDX-License-Identifier` at the top with one of the
 
 Continuous integration is hosted at https://ci.arvados.org/
 
-Currently, external contributers cannot trigger builds.  We are investigating integration with Github pull requests for the future.
+Currently, external contributors cannot trigger builds.  We are investigating integration with Github pull requests for the future.
 
 [![Build Status](https://ci.arvados.org/buildStatus/icon?job=run-tests)](https://ci.arvados.org/job/run-tests/)
 
index 9a32de0d0b35ba335b9890ce918188ecf987b866..d6a3bb4c10eaaff84e6fb337d1fa32538800654f 100644 (file)
@@ -18,7 +18,7 @@ A version will be saved when one of the following conditions is true:
 
 One is by "configuring (system-wide) the collection's idle time":{{site.baseurl}}/admin/collection-versioning.html. This idle time is checked against the @modified_at@ attribute so that the version is saved when one or more of the previously enumerated attributes get updated and the @modified_at@ is at least at the configured idle time in the past. This way, a frequently updated collection won't create lots of version records that may not be useful.
 
-The other way to trigger a version save, is by setting @preserve_version@ to @true@ on the current version collection record: this ensures that the current state will be preserved as a version the next time it gets updated.
+The other way to trigger a version save, is by setting @preserve_version@ to @true@ on the current version collection record: this ensures that the current state will be preserved as a version the next time it gets updated. This includes either creating a new collection or updating a preexisting one. In the case of using @preserve_version = true@ on a collection's create call, the new record state will be preserved as a snapshot on the next update.
 
 h3. Collection's past versions behavior & limitations
 
index d224924a24914a0d88af34330b34cff52ed3ec2e..4c4e341f5a34c8608e554f407aab4869da88a8f6 100644 (file)
@@ -96,7 +96,7 @@ var whitelist = map[string]bool{
        "Collections.BlobTrashCheckInterval":                  false,
        "Collections.BlobTrashConcurrency":                    false,
        "Collections.BlobTrashLifetime":                       false,
-       "Collections.CollectionVersioning":                    false,
+       "Collections.CollectionVersioning":                    true,
        "Collections.DefaultReplication":                      true,
        "Collections.DefaultTrashLifetime":                    true,
        "Collections.ForwardSlashNameSubstitution":            true,
index e39fdd8d9f2db8259a210649d279b08f163d054e..f034ca5ab2f1ed82a385431fb1e32acea7c17fc7 100644 (file)
@@ -39,8 +39,8 @@ setup(name='arvados-cwl-runner',
       # file to determine what version of cwltool and schema-salad to
       # build.
       install_requires=[
-          'cwltool==3.1.20211020155521',
-          'schema-salad==8.2.20211020114435',
+          'cwltool==3.1.20211107152837',
+          'schema-salad==8.2.20211116214159',
           'arvados-python-client{}'.format(pysdk_dep),
           'setuptools',
           'ciso8601 >= 2.0.0',
index d03265ca44b1b6e886e57d455dadcb7a613d0b6e..55be40fa04fc0f7a46a5de5a239464547f81df4b 100644 (file)
@@ -1546,7 +1546,8 @@ class Collection(RichCollectionBase):
              storage_classes=None,
              trash_at=None,
              merge=True,
-             num_retries=None):
+             num_retries=None,
+             preserve_version=False):
         """Save collection to an existing collection record.
 
         Commit pending buffer blocks to Keep, merge with remote record (if
@@ -1576,6 +1577,13 @@ class Collection(RichCollectionBase):
         :num_retries:
           Retry count on API calls (if None,  use the collection default)
 
+        :preserve_version:
+          If True, indicate that the collection content being saved right now
+          should be preserved in a version snapshot if the collection record is
+          updated in the future. Requires that the API server has
+          Collections.CollectionVersioning enabled, if not, setting this will
+          raise an exception.
+
         """
         if properties and type(properties) is not dict:
             raise errors.ArgumentError("properties must be dictionary type.")
@@ -1588,6 +1596,9 @@ class Collection(RichCollectionBase):
         if trash_at and type(trash_at) is not datetime.datetime:
             raise errors.ArgumentError("trash_at must be datetime type.")
 
+        if preserve_version and not self._my_api().config()['Collections'].get('CollectionVersioning', False):
+            raise errors.ArgumentError("preserve_version is not supported when CollectionVersioning is not enabled.")
+
         body={}
         if properties:
             body["properties"] = properties
@@ -1596,6 +1607,8 @@ class Collection(RichCollectionBase):
         if trash_at:
             t = trash_at.strftime("%Y-%m-%dT%H:%M:%S.%fZ")
             body["trash_at"] = t
+        if preserve_version:
+            body["preserve_version"] = preserve_version
 
         if not self.committed():
             if self._has_remote_blocks:
@@ -1641,7 +1654,8 @@ class Collection(RichCollectionBase):
                  storage_classes=None,
                  trash_at=None,
                  ensure_unique_name=False,
-                 num_retries=None):
+                 num_retries=None,
+                 preserve_version=False):
         """Save collection to a new collection record.
 
         Commit pending buffer blocks to Keep and, when create_collection_record
@@ -1680,6 +1694,13 @@ class Collection(RichCollectionBase):
         :num_retries:
           Retry count on API calls (if None,  use the collection default)
 
+        :preserve_version:
+          If True, indicate that the collection content being saved right now
+          should be preserved in a version snapshot if the collection record is
+          updated in the future. Requires that the API server has
+          Collections.CollectionVersioning enabled, if not, setting this will
+          raise an exception.
+
         """
         if properties and type(properties) is not dict:
             raise errors.ArgumentError("properties must be dictionary type.")
@@ -1690,6 +1711,9 @@ class Collection(RichCollectionBase):
         if trash_at and type(trash_at) is not datetime.datetime:
             raise errors.ArgumentError("trash_at must be datetime type.")
 
+        if preserve_version and not self._my_api().config()['Collections'].get('CollectionVersioning', False):
+            raise errors.ArgumentError("preserve_version is not supported when CollectionVersioning is not enabled.")
+
         if self._has_remote_blocks:
             # Copy any remote blocks to the local cluster.
             self._copy_remote_blocks(remote_blocks={})
@@ -1718,6 +1742,8 @@ class Collection(RichCollectionBase):
             if trash_at:
                 t = trash_at.strftime("%Y-%m-%dT%H:%M:%S.%fZ")
                 body["trash_at"] = t
+            if preserve_version:
+                body["preserve_version"] = preserve_version
 
             self._remember_api_response(self._my_api().collections().create(ensure_unique_name=ensure_unique_name, body=body).execute(num_retries=num_retries))
             text = self._api_response["manifest_text"]
index 8d637303b466f90712b9ef4b44846277d5af6ed0..f82d44ab6033072915044e0514663ad9e1890025 100644 (file)
@@ -50,7 +50,7 @@ setup(name='arvados-python-client',
           'future',
           'google-api-python-client >=1.6.2, <2',
           'google-auth<2',
-          'httplib2 >=0.9.2',
+          'httplib2 >=0.9.2, <0.20.2',
           'pycurl >=7.19.5.1',
           'ruamel.yaml >=0.15.54, <0.17.11',
           'setuptools',
index 6d2643a967ef70374f2ff222c19bb67917ba5a0e..f9178325091f94ebb97e5f46fba02903adf99104 100644 (file)
@@ -791,6 +791,7 @@ def setup_config():
                     "UserProfileNotificationAddress": "arvados@example.com",
                 },
                 "Collections": {
+                    "CollectionVersioning": True,
                     "BlobSigningKey": "zfhgfenhffzltr9dixws36j1yhksjoll2grmku38mi7yxd66h5j4q9w4jzanezacp8s6q0ro3hxakfye02152hncy6zml2ed0uc",
                     "TrustAllContent": False,
                     "ForwardSlashNameSubstitution": "/",
index f821ff952f7a45f913538c890ffc40d397b04ada..a43e0d40dfe7ed48f5477689d3623afefe952ba3 100644 (file)
@@ -1360,6 +1360,25 @@ class NewCollectionTestCaseWithServersAndTokens(run_test_server.TestCaseWithServ
 
 
 class NewCollectionTestCaseWithServers(run_test_server.TestCaseWithServers):
+    def test_preserve_version_on_save(self):
+        c = Collection()
+        c.save_new(preserve_version=True)
+        coll_record = arvados.api().collections().get(uuid=c.manifest_locator()).execute()
+        self.assertEqual(coll_record['version'], 1)
+        self.assertEqual(coll_record['preserve_version'], True)
+        with c.open("foo.txt", "wb") as foo:
+            foo.write(b"foo")
+        c.save(preserve_version=True)
+        coll_record = arvados.api().collections().get(uuid=c.manifest_locator()).execute()
+        self.assertEqual(coll_record['version'], 2)
+        self.assertEqual(coll_record['preserve_version'], True)
+        with c.open("bar.txt", "wb") as foo:
+            foo.write(b"bar")
+        c.save(preserve_version=False)
+        coll_record = arvados.api().collections().get(uuid=c.manifest_locator()).execute()
+        self.assertEqual(coll_record['version'], 3)
+        self.assertEqual(coll_record['preserve_version'], False)
+
     def test_get_manifest_text_only_committed(self):
         c = Collection()
         with c.open("count.txt", "wb") as f:
index 8d15bb1c5062a0215a04f60fb9749cd7ad7e35d1..7fbb86c010c01df4d08cbde40c434c7f8b9063ae 100644 (file)
@@ -10,6 +10,8 @@ class Arvados::V1::GroupsController < ApplicationController
   skip_before_action :find_object_by_uuid, only: :shared
   skip_before_action :render_404_if_no_object, only: :shared
 
+  TRASHABLE_CLASSES = ['project']
+
   def self._index_requires_parameters
     (super rescue {}).
       merge({
@@ -99,6 +101,15 @@ class Arvados::V1::GroupsController < ApplicationController
     end
   end
 
+  def destroy
+    if !TRASHABLE_CLASSES.include?(@object.group_class)
+      return @object.destroy
+      show
+    else
+      super # Calls destroy from TrashableController module
+    end
+  end
+
   def render_404_if_no_object
     if params[:action] == 'contents'
       if !params[:uuid]
@@ -351,8 +362,6 @@ class Arvados::V1::GroupsController < ApplicationController
     @offset = offset_all
   end
 
-  protected
-
   def exclude_home objectlist, klass
     # select records that are readable by current user AND
     #   the owner_uuid is a user (but not the current user) OR
index 02a4ce96632d2962b830a720fbe3621458e79fb8..4dbccc5eb24edd07cdb09e6f2e400faa9c4f0c81 100644 (file)
@@ -538,6 +538,21 @@ class Arvados::V1::GroupsControllerTest < ActionController::TestCase
     assert_includes(owners, groups(:asubproject).uuid)
   end
 
+  [:afiltergroup, :private_role].each do |grp|
+    test "delete non-project group #{grp}" do
+      authorize_with :admin
+      assert_not_nil Group.find_by_uuid(groups(grp).uuid)
+      assert !Group.find_by_uuid(groups(grp).uuid).is_trashed
+      post :destroy, params: {
+            id: groups(grp).uuid,
+            format: :json,
+          }
+      assert_response :success
+      # Should not be trashed
+      assert_nil Group.find_by_uuid(groups(grp).uuid)
+    end
+  end
+
   ### trashed project tests ###
 
   #
index f076ccf18419675499e12eed0e3d017824af8e57..46f4db4095bfb286c82f4f07a988c16cee5ebe63 100644 (file)
@@ -359,47 +359,53 @@ var blockFileRe = regexp.MustCompile(`^[0-9a-f]{32}$`)
 //     e4de7a2810f5554cd39b36d8ddb132ff+67108864 1388701136
 //
 func (v *UnixVolume) IndexTo(prefix string, w io.Writer) error {
-       var lastErr error
        rootdir, err := v.os.Open(v.Root)
        if err != nil {
                return err
        }
-       defer rootdir.Close()
        v.os.stats.TickOps("readdir")
        v.os.stats.Tick(&v.os.stats.ReaddirOps)
-       for {
-               names, err := rootdir.Readdirnames(1)
-               if err == io.EOF {
-                       return lastErr
-               } else if err != nil {
-                       return err
-               }
-               if !strings.HasPrefix(names[0], prefix) && !strings.HasPrefix(prefix, names[0]) {
+       subdirs, err := rootdir.Readdirnames(-1)
+       rootdir.Close()
+       if err != nil {
+               return err
+       }
+       for _, subdir := range subdirs {
+               if !strings.HasPrefix(subdir, prefix) && !strings.HasPrefix(prefix, subdir) {
                        // prefix excludes all blocks stored in this dir
                        continue
                }
-               if !blockDirRe.MatchString(names[0]) {
+               if !blockDirRe.MatchString(subdir) {
                        continue
                }
-               blockdirpath := filepath.Join(v.Root, names[0])
+               blockdirpath := filepath.Join(v.Root, subdir)
                blockdir, err := v.os.Open(blockdirpath)
                if err != nil {
                        v.logger.WithError(err).Errorf("error reading %q", blockdirpath)
-                       lastErr = fmt.Errorf("error reading %q: %s", blockdirpath, err)
-                       continue
+                       return fmt.Errorf("error reading %q: %s", blockdirpath, err)
                }
                v.os.stats.TickOps("readdir")
                v.os.stats.Tick(&v.os.stats.ReaddirOps)
-               for {
-                       fileInfo, err := blockdir.Readdir(1)
-                       if err == io.EOF {
-                               break
+               // ReadDir() (compared to Readdir(), which returns
+               // FileInfo structs) helps complete the sequence of
+               // readdirent calls as quickly as possible, reducing
+               // the likelihood of NFS EBADCOOKIE (523) errors.
+               dirents, err := blockdir.ReadDir(-1)
+               blockdir.Close()
+               if err != nil {
+                       v.logger.WithError(err).Errorf("error reading %q", blockdirpath)
+                       return fmt.Errorf("error reading %q: %s", blockdirpath, err)
+               }
+               for _, dirent := range dirents {
+                       fileInfo, err := dirent.Info()
+                       if os.IsNotExist(err) {
+                               // File disappeared between ReadDir() and now
+                               continue
                        } else if err != nil {
-                               v.logger.WithError(err).Errorf("error reading %q", blockdirpath)
-                               lastErr = fmt.Errorf("error reading %q: %s", blockdirpath, err)
-                               break
+                               v.logger.WithError(err).Errorf("error getting FileInfo for %q in %q", dirent.Name(), blockdirpath)
+                               return err
                        }
-                       name := fileInfo[0].Name()
+                       name := fileInfo.Name()
                        if !strings.HasPrefix(name, prefix) {
                                continue
                        }
@@ -408,16 +414,15 @@ func (v *UnixVolume) IndexTo(prefix string, w io.Writer) error {
                        }
                        _, err = fmt.Fprint(w,
                                name,
-                               "+", fileInfo[0].Size(),
-                               " ", fileInfo[0].ModTime().UnixNano(),
+                               "+", fileInfo.Size(),
+                               " ", fileInfo.ModTime().UnixNano(),
                                "\n")
                        if err != nil {
-                               blockdir.Close()
                                return fmt.Errorf("error writing: %s", err)
                        }
                }
-               blockdir.Close()
        }
+       return nil
 }
 
 // Trash trashes the block data from the unix storage