Node Recovery

1. Node Recovery • Day 3, Session 6 • MetaArchive Distributed Digital Preservation Workshop Presented by Chris Helms

2. Session Learning Objectives

3. Restoring a lost cache • Hardware/Software fault recovery • Archival Units • Archival Unit Configuration

4. Restoring a lost primary node • Preparing for Configuration Node Recovery • Recovering the Title Database • Recovering the Keystore • Reconfiguring the Nodes in the Network

5. Restoring data from the cache • Viewing Cached Content • Using your LOCKSS node as a cache proxy • Proxy configuration options • Cached AU retrieval and proxy testing

6. Restoring a Lost Cache • By design, LOCKSS stores harvested material in a cache directory. LOCKSS will maintain the cache’s integrity through peer polling and voting with other members of your LOCKSS private network. Occasionally an outside occurrence such as a power failure, problematic hardware, or a piece of errant coding can corrupt a file system leaving the LOCKSS cache in an unusable state.

7. Archival Units • Archival Units may be likened to a run of journals or a web-based collection containing images, videos, and text. Long term access to these AU’s is insured via LOCKSS harvesting and the polling of loyal peers. The concept of having loyal peers insures that data may be gained from other peer nodes if the original content is no longer available from the publisher.

8. Archival Configuration Adding Titles (manual)

9. Title Selection

10. Archival Configuration Adding Titles (backup file) Default cache config file = BatchAuConfig

11. Restoring a Lost Primary Node • In order for LOCKSS to harvest and preserve data, certain configuration files must remain available via a primary node. These are the title database and public keystore. As a convenience, these files are stored within the conspectus database and are readily available from any participating LOCKSS node.

12. Recovering the Title Database

13. Recovering the Keystore

14. Reconfiguring the Nodes in the Network Creation and setup of the new web root directory and Apache configuration files. Modify the following line: Apache graceful restart

15. Update the lockss.xml file

16. Run hostconfig on all member nodes and modify the “Configuration URL:”

17. Restart LOCKSS • /etc/init.d/lockss stop • /etc/init.d/lockss start

18. Restoring Data From the Cache • One of the key benefits of participating in a LOCKSS private network is having reliable, long-term access to your material. When available access to the original content is no longer available, the LOCKSS cache proxy server can be used to serve its harvested content. This harvested content may be pulled from any node participating in your LOCKSS private network. As we have seen in the previous section, harvested content may span archival units, key configuration files, and plugins.

19. Viewing cached content

23. Using your LOCKSS node as a cache proxy • A LOCKSS node runs a cache proxy that will serve cached content if the original content is no longer available. If the original content is still available, the LOCKSS node will proxy the request to the original site.

24. Proxy access control

25. Proxy Options Notes: 1. The audit proxy serves only cached content, and never forwards requests to the publisher or any other sites. By configuring a browser to proxy to this port, you can view the content stored on the cache. All requests for content not on the cache will return a “404 Not Found” error. 2. The ICP server responds to queries sent to other proxies and caches to let them know if this cache has the content they are looking for. This is useful if you are integrating this cache into an exiting network structure with other proxies and cache that support ICP.

26. Proxy Info

27. Excerpt from a PAC file:

28. Adding a PAC file to Mozilla Firefox 1. Windows: Tools/Options MAC OS X: File/Preferences 2. Advanced/Network/Settings

29. Questions ?