openstack/openstack-ansible
Code Issues Proposed changes
00f276b9b9
openstack-ansible/doc/source/admin/scaling-swift.rst
Jean-Philippe Evrard 14ae5a8b64 [Docs] Promote scale the environment 
Scale the environment deserves its own chapter.
On top of that, the managing swift section is
about how scaling is done with swift clusters
(Multi-Region handling), and therefore should
be moved into scaling the environment.

Change-Id: I3a37bfefb53e229278c7ec08a8acde63614949ed
2018-02-20 07:57:05 +00:00

2.5 KiB
Raw Blame History

Accessibility for multi-region Object Storage

In multi-region Object Storage utilizing separate database backends, objects are retrievable from an alternate location if the default_project_id for a user in the keystone database is the same across each database backend.

Important

It is recommended to perform the following steps before a failure occurs to avoid having to dump and restore the database.

If a failure does occur, follow these steps to restore the database from the Primary (failed) Region:

  1. Record the Primary Region output of the default_project_id for the specified user from the user table in the keystone database:

    Note

    The user is admin in this example.

    # mysql -e "SELECT default_project_id from keystone.user WHERE \
  name='admin';"

+----------------------------------+
| default_project_id               |
+----------------------------------+
| 76ef6df109744a03b64ffaad2a7cf504 |
+-----------------—————————————————+

  2. Record the Secondary Region output of the default_project_id for the specified user from the user table in the keystone database:

    # mysql -e "SELECT default_project_id from keystone.user WHERE \
  name='admin';"

+----------------------------------+
| default_project_id               |
+----------------------------------+
| 69c46f8ad1cf4a058aa76640985c     |
+----------------------------------+

  3. In the Secondary Region, update the references to the project_id to match the ID from the Primary Region:

    # export PRIMARY_REGION_TENANT_ID="76ef6df109744a03b64ffaad2a7cf504"
# export SECONDARY_REGION_TENANT_ID="69c46f8ad1cf4a058aa76640985c"

# mysql -e "UPDATE keystone.assignment set \
target_id='${PRIMARY_REGION_TENANT_ID}' \
WHERE target_id='${SECONDARY_REGION_TENANT_ID}';"

# mysql -e "UPDATE keystone.user set \
default_project_id='${PRIMARY_REGION_TENANT_ID}' WHERE \
default_project_id='${SECONDARY_REGION_TENANT_ID}';"

# mysql -e "UPDATE keystone.project set \
id='${PRIMARY_REGION_TENANT_ID}' WHERE \
id='${SECONDARY_REGION_TENANT_ID}';"

The user in the Secondary Region now has access to objects PUT in the Primary Region. The Secondary Region can PUT objects accessible by the user in the Primary Region.