es: backup and restore

This PR provides CLI commands to backup and restore elasticsearch indices and data.

* Installs CLI commands in `setup.py`.
* Creates CLI commands for creating a repository, creating a snapshot and restoring a snapshot.
* Mount a temp path in elasticsearch docker service and add the `path.repo` configuration to make snapshot working.

Co-Authored-by: Sébastien Délèze <sebastien.deleze@rero.ch>

docker-services.yml

@@ -81,6 +81,7 @@ services:
       - "ES_JAVA_OPTS=-Xms512m -Xmx512m"
       - "discovery.type=single-node"
       - "indices.query.bool.max_clause_count=3000"
+      - "path.repo=/var/tmp"
     ulimits:
       memlock:
         soft: -1
@@ -92,6 +93,8 @@ services:
     ports:
       - "9200:9200"
       - "9300:9300"
+    volumes:
+      - /var/tmp
   kibana:
     image: docker.elastic.co/kibana/kibana-oss:7.9.1
     environment:

setup.py

@@ -53,7 +53,8 @@
             'documents = sonar.modules.documents.cli.documents:documents',
             'oaiharvester = \
                 sonar.modules.documents.cli.oaiharvester:oaiharvester',
-            'utils = sonar.modules.cli:utils'
+            'utils = sonar.modules.cli:utils',
+            'es = sonar.elasticsearch.cli:es'
         ],
         'invenio_base.apps': [
             'sonar = sonar.modules:Sonar',

sonar/elasticsearch/__init__.py

@@ -0,0 +1,18 @@
+# -*- coding: utf-8 -*-
+#
+# Swiss Open Access Repository
+# Copyright (C) 2019 RERO
+#
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU Affero General Public License as published by
+# the Free Software Foundation, version 3 of the License.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU Affero General Public License for more details.
+#
+# You should have received a copy of the GNU Affero General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+"""Interactions with Elasticsearch."""

sonar/elasticsearch/cli.py

@@ -0,0 +1,139 @@
+# -*- coding: utf-8 -*-
+#
+# Swiss Open Access Repository
+# Copyright (C) 2019 RERO
+#
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU Affero General Public License as published by
+# the Free Software Foundation, version 3 of the License.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU Affero General Public License for more details.
+#
+# You should have received a copy of the GNU Affero General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+"""CLI for Elasticsearch."""
+
+import datetime
+
+import click
+from flask.cli import with_appcontext
+from invenio_search import current_search_client
+
+
+def abort_if_false(ctx, param, value):
+    """Abort command is value is False."""
+    if not value:
+        ctx.abort()
+
+
+@click.group()
+def es():
+    """Commands for ES."""
+
+
+@click.group()
+def snapshot():
+    """Snapshot commands."""
+
+
+@click.command()
+@click.option('--repository', 'repository', default='backup')
+@with_appcontext
+def create_repository(repository):
+    """Create repository for snapshot.
+
+    :param repository: Repository name.
+    """
+    click.secho('Create a repository for snapshots')
+
+    try:
+        current_search_client.snapshot.create_repository(
+            repository, {
+                'type': 'fs',
+                'settings': {
+                    'location': repository
+                }
+            })
+        click.secho('Done', fg='green')
+    except Exception as exception:
+        click.secho(str(exception), fg='red')
+
+
+@click.command()
+@click.option('--repository', 'repository', default='backup')
+@click.option('--name', 'name')
+@click.option('--wait/--no-wait', default=False)
+@with_appcontext
+def backup(repository, name, wait):
+    """Backup elasticsearch data.
+
+    :param repository: Repository name.
+    :param name: Name of the snapshot.
+    :param wait: Wait for completion.
+    """
+    click.secho('Backup elasticsearch data')
+
+    # If no name, create a snapshot with the current date
+    if not name:
+        name = 'snapshot-{date}'.format(
+            date=datetime.date.today().strftime('%Y-%m-%d'))
+
+        click.secho('Create a snapshot with name {name}'.format(name=name))
+
+    try:
+        # Remove old backup with the same name
+        try:
+            current_search_client.snapshot.delete(repository, name)
+        except Exception:
+            pass
+
+        # Backup data
+        current_search_client.snapshot.create(repository,
+                                              name,
+                                              wait_for_completion=wait)
+
+        click.secho('Done', fg='green')
+    except Exception as exception:
+        click.secho(str(exception), fg='red')
+
+
+@click.command()
+@click.option('--repository', 'repository', default='backup')
+@click.option('--name', 'name', required=True)
+@click.option('--wait/--no-wait', default=False)
+@click.option('--yes-i-know',
+              is_flag=True,
+              callback=abort_if_false,
+              expose_value=False,
+              prompt='Do you really want to restore a snapshot?')
+@with_appcontext
+def restore(repository, name, wait):
+    """Restore elasticsearch data.
+
+    :param repository: Repository name.
+    :param name: Name of the snapshot.
+    :param wait: Wait for completion.
+    """
+    click.secho('Restore elasticsearch data')
+    try:
+        # Remove all indices
+        current_search_client.indices.delete('_all')
+
+        # Restore the snapshot
+        current_search_client.snapshot.restore(repository,
+                                               name,
+                                               wait_for_completion=wait)
+
+        click.secho('Done', fg='green')
+    except Exception as exception:
+        click.secho(str(exception), fg='red')
+
+
+snapshot.add_command(create_repository)
+snapshot.add_command(backup)
+snapshot.add_command(restore)
+es.add_command(snapshot)

tests/unit/elasticsearch/test_elasticsearch_cli.py

@@ -0,0 +1,99 @@
+# -*- coding: utf-8 -*-
+#
+# Swiss Open Access Repository
+# Copyright (C) 2019 RERO
+#
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU Affero General Public License as published by
+# the Free Software Foundation, version 3 of the License.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU Affero General Public License for more details.
+#
+# You should have received a copy of the GNU Affero General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+"""Test Elasticsearch cli commands."""
+
+import datetime
+
+from click.testing import CliRunner
+from invenio_search import current_search_client
+from mock import patch
+
+from sonar.elasticsearch.cli import backup, create_repository, restore
+
+
+def test_create_repository(app, script_info):
+    """Test create repository."""
+    runner = CliRunner()
+
+    # OK
+    result = runner.invoke(create_repository, obj=script_info)
+    assert result.output == 'Create a repository for snapshots\nDone\n'
+
+    # Repository creation failed
+    with patch(
+            'invenio_search.current_search_client.snapshot.create_repository',
+            side_effect=Exception('Mocked error')):
+        result = runner.invoke(create_repository, obj=script_info)
+        assert result.output == 'Create a repository for snapshots\nMocked ' \
+            'error\n'
+
+
+def test_backup(app, script_info):
+    """Test backup."""
+    runner = CliRunner()
+
+    # OK
+    current_search_client.snapshot.create_repository('backup', {
+        'type': 'fs',
+        'settings': {
+            'location': 'backup'
+        }
+    })
+    result = runner.invoke(backup, ['--name', 'test'], obj=script_info)
+    assert result.output == 'Backup elasticsearch data\nDone\n'
+
+    # Snapshot with no name
+    result = runner.invoke(backup, obj=script_info)
+    assert 'snapshot-{date}'.format(
+        date=datetime.date.today().strftime('%Y-%m-%d')) in result.output
+
+    # Not existing repository
+    current_search_client.snapshot.delete('backup', 'test')
+    current_search_client.snapshot.delete_repository('backup')
+    result = runner.invoke(backup, obj=script_info)
+    assert 'repository_missing_exception' in result.output
+
+
+def test_restore(app, script_info):
+    """Test restore."""
+    runner = CliRunner()
+
+    current_search_client.snapshot.create_repository('backup', {
+        'type': 'fs',
+        'settings': {
+            'location': 'backup'
+        }
+    })
+    result = runner.invoke(backup, ['--name', 'test', '--wait'],
+                           obj=script_info)
+
+    # OK
+    result = runner.invoke(restore, ['--name', 'test', '--yes-i-know'],
+                           obj=script_info)
+    assert result.output == 'Restore elasticsearch data\nDone\n'
+
+    # Unexisting snapshot
+    result = runner.invoke(restore, ['--name', 'unexisting', '--yes-i-know'],
+                           obj=script_info)
+    assert 'snapshot does not exist' in result.output
+
+    # Aborting
+    result = runner.invoke(restore, ['--name', 'test'],
+                           obj=script_info,
+                           input='N')
+    assert 'Aborted!\n' in result.output