Skip to end of metadata
Go to start of metadata

You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 20 Next »

Here is the current backup strategy as a diagram:

Recovery Options

Fedora:

S3:

Currently we are using the s3 sync tool (akin to rsync for S3) to pull over key fedora data into the chf-hydra-backup (https://s3.console.aws.amazon.com/s3/buckets/chf-hydra-backup) bucket. This is a slight misnomer as it handles backups for ArchivesSpace as well now, but Fedora data is pulled over into s3 key FedoraBackup (https://s3.console.aws.amazon.com/s3/buckets/chf-hydra-backup/FedoraBackup/?region=us-east-1&tab=overview) which contains all Fedora binary data.

PGSql (https://s3.console.aws.amazon.com/s3/buckets/chf-hydra-backup/PGSql/ ) contains the Fedora Postgres database fcrepo_backup)

Both https://s3.console.aws.amazon.com/s3/buckets/chf-hydra-backup/FedoraBackup/ and https://s3.console.aws.amazon.com/s3/object/chf-hydra-backup/PGSql/fcrepo_backup.sql are needed to do a full restore.

Note: As a reminder, while S3's visual interface uses folders, those locations are actually just the first step in a path of individual block stored objects. Folders do not exist in S3.

How to restore Fedora:

  1. Stop Tomcat
  2. Download the Postgres database fcrepo_backup.sql to an arbitrary location on the Fedora machine.

  3. Fedora might still have active connections to postgres, so run a postgres restart to kill them: sudo service postgresql restart .

  4. Import the database: psql fcrepo < fcrepo_backup.sql
    1. If the database already exists, such as when you are running a sync, you will want to drop the existing database and then run the command.
  5. Check that the user trilby has permissions to access and use the newly made fcrepo database.
  6. Delete the existing folder(s) inside /opt/fedora-data (This step is not always required but makes it simpler)
  7. Using screen or tmux start an aws s3 sync to copy all the data over in the FedoraBackup "folder" to /opt/fedora-data : aws s3 sync s3://chf-hydra-backup/FedoraBackup /opt/fedora-data/
  8. Wait a while for all the data (>800 GB) to copy over.
  9. Run chown -R tomcat8:tomcat8 /opt/fedora-data to give ownership on the new files to the tomcat user so Fedora can access them.
  10. Restart Tomcat: sudo service tomcat8 restart OR sudo systemctl tomcat8 restart
  11. This will restore the Fedora database.  Current cost estimates (2/18) are about $.10 to do this restore.

How to restore users:

  1. Go to S3 and download the postgres backup files to an arbitrary location on the app server.
  2. Stop Apache
  3. Restart the postgres service (see above). This should remove the default connection to the Sufia database that Sufia has when it's running, so you can change it.
  4. In Postgres, delete the automatically generated chf_hydra database as follows:
    1. Log in via psql -u postgres
      1. The postgres account password is in ansible-vault (groupvars/all)
    2. Run: DROP DATABASE chf_hydra;
    3. Run: CREATE DATABASE chf_hydra;
  5. Then import the downloaded database
    1. Either:
      1. pg_restore -d chf_hydra -U postgres chf_hydra.dump
      2. psql chf_hydra < chf_hydra_dump.sql
  6. Then set permissions
    1. psql -U postgres
    2. GRANT Create,Connect,Temporary ON DATABASE chf_hydra TO chf_pg_hydra;
  7. You may now restart postgres and Apache2. systemctl restart apache2

Note: the minter is now part of postgres: no need to take any extra steps. just restore the chf_hydra database for the users to app and the minter will be restored.

How to restore redis:

Redis keeps a database in memory which handles the transaction record data such as the history of edits on a record. It does not contain the actual data, simply the timeline of changes. Losing this causes the history of object edits to be lost, but the objects themselves will be fine.

  1. Download redis-dump/dump.rdb to an arbitrary location on the app server.
  2. It must be changed to be owned by the redis user as follows:
    1. sudo chmod -R redis:redis filename
  3. Stop the redis server as follows:
    1. sudo service redis-server stop
  4. Move redis-dump.rdb to /var/lib/redis/dump.rdb . This will overwrite the existing file there called dump.rdb
  5. Restart redis
    1. sudo service redis-server start
  6. When starting, redis will read the .rdb dump file and copy that data back into the in-memory database.

Indexing:

The index is being backed up to speed up the time to recovery for DR or migrations. If you cannot access it, a manual reindex can be done with the instructions in Application administration. This process takes at least one business day, so is not recommended versus rebuilding from the backup.

  1. In the chf-hydra-backup, pull down the solr-backup.tar.gz file under Solr to the Solr server.
  2. Extract the archive
  3. Use the solr restore commands at Application administration 

Costs

A quick cost analysis has restoration costing $30-35 dollars, this is as of 6/11/2018 with approximately 1 TB of data. Approximately 66% of the cost was due to inter-region transfer fees (moving data from US-WEST to US-EAST). The rest is standard LIST, GET, and related fees.

Scihist_digicoll Backup and Recovery


Digital Collection Recovery Overview:

Our new digital collection application (scihist_digcol) has slightly different preservation needs than the current Samvera application. However, our institutional goals remain the same. While we will be making technical changes to preservation methods, our backup and recovery goals are still linked to the following key points in order of importance:

  1. Provide backups of all original files to they can be recovered in case of natural disaster, data corruption, user error, or malicious computer activity.

  2. Allow the Institute to recover from data loss or outages in a reasonable rate of time.

  3. Adhere to OAIS model rules when possible.

The following classifications of recovery may prove helpful

  • Partial public, partial staff recovery: The public has access to limited functionality, staff has access to a limited set of functions but certain functions may not be used. This is considered an incomplete recovery.

  • Full public, partial staff recovery: The public can use the site normally, staff has access to a limited set of functions but certain functions may not be used. This is also an incomplete recovery but time sensitivity is reduced as public users are not impacted. Staff recovery times should be minimized, but public use takes priority.

  • Full public, full staff recovery: A total recovery.

Then there are levels of disaster/data loss

  • Minimal Data Loss: In this case a small subset of data is lost, general site functionality is unaffected.

  • Minimal Data Inaccessibility: A small subset of data is temporarily inaccessible, general site functionality is unaffected.

  • Major Data Inaccessibility: Data is not lost permanently, but our ability to access the data is compromised. In this case general site functionality is affected.

  • Major Data Loss: A large amount of data is lost, and general site functionality is impacted.

Generally speaking, public recovery is the higher initial priority since it impacts more people, contacting staff about outages is easy, and public outages affect the perception of the institute.

Inaccessibility is also different from data loss, though they share certain characteristics. In both cases a solution is to have additional copies of data, but for inaccessibility it is so they can be used as a temporary source of data until the outage is resolved. Handling data inaccessibility requires that the secondary source of data be similarly structured to minimize the time to switch over. With data loss the backups can be in any format as the intent is that a new source of data will be built from the backups.

Our data can be broken into two categories, one is data that is potentially irrecoverable. This includes our original binary files (images, audio, other) and the metadata about them (a SQL database). The other data is restorable but needed for normal site operation but takes significant time to restore, such as the derived download files and viewer files. The second set of data may be worth backing up to shorten recovery times for public users when data is lost.

As an estimate, our cost to hold extra backups for our current scihistcoll staging environment costs less than $2 a month out of a total $70.07 spent on data storage inclusive of data transfer and storage. While a production environment will have slightly higher cost ratios, it should not be massively higher. Thus by spending an additional 2-3% cost on S3, we can mount a full public recovery in an afternoon from a massive failure of our entire infrastructure. While we currently are not backing up our viewer tiles, an examination of our old application shows the cost for production averages around 5 dollars for storage. Adding a second copy of the viewer files should roughly double the cost, with a slight reduction for less use, so will add another 5 dollars to the cost, so for about $7-12 dollars a month we can be widely covered for data inaccessibility or other failures of S3 in a specific region. While it is hard to get specific details, there have been multiple outages or issues in a region whose duration lasted over an hour, and at least one major outage in the last two years lasting around 6 hours. Assuming about 8 hours of problems every two years, we can estimate that a rough cost of $36/hour of outage spent to avoid being down. Shorter outages may not be worth the difficulty of switching over.

In cases of small scale data loss, such as corrupted files or user error, the application will be working fine but a limited set of data will have a problem. In these cases we can locate the problematic data and use a backup copy to restore any damaged original files or use versioning to restore an earlier version of the file. Derived files can either be regenerated or copied from backups as well. This is the most common expected use case at requires only that we keep versions of our original files and backups of files in different locations (another S3 bucket and an on-site copy).

In cases of broad data loss most or all of the data is rendered unavailable. In these cases we will suffer a loss of service until we can recover the data. This can be thought of as two recoveries: one is to get the digital collection site back as soon as possible for the public and the second is to restore all functionality. Getting the site back for the public is our primary concern, so as noted above for outages we have a few methods to speed up recovery at an additional cost to our backup costs. Both the derivative and original files are backed up to a region on the West Coast in S3 with the same configuration details that our files in US-EAST (the original originals and original derivatives) have. We can recover public access by using these backup files directly while we spend more time working on a full recovery for staff functionality. Due to current setups we would not want staff to add new works, but this allows us to rapidly restore public facing access to our site should the normal data sources be unavailable. A longer process allows us to restore the data back to the original locations while leaving public access up, once the data is restored to its original place full staff functionality will likewise be restored.

Finally in cases of issues affecting all on-line storage systems, another copy of the data is held on our in-house storage system so that we can potentially recover data even in case of a full loss of AWS. This only holds the original files and all other aspects will need to be rebuild, a process that can take days in addition to the time taken to upload the files again. Using the local backup means recovery could take a week or more. This is not paid for/managed by our team as Institute IT handles these systems.

Technical Notes

Kithe currently (March 2019) has a small set of data to be handled for recovery.

  • A postgres database which contains user data and item metadata

  • Original binary files

  • Derivative Files

The first two are the ones that are needed. If they were lost, the derivatives can be easily recreated based on the first two, but it would take over a day. Thus, we keep a backup anyway because it's cheap and minimizes downtime.

  • The postgres database is to be backed up to S3, with a version history of the last 30 versions of the file representing roughly a month of backups.
  • The binary files are replicated via S3 replication to a second location in US-WEST rather than US-EAST in case of outages. (During development, as of summer 2019, these binaries are stored at https://s3.console.aws.amazon.com/s3/buckets/scihi-kithe-stage-originals .) When, as part of launching the site, we actually switch over to production, these will also be backed up over to local on-site storage. The binary files are also versioned and prior versions are held for 30 days before being cleared away to reduce storage costs. This offers a month period to revert a file back if something is damaged.
  • The derivative files will also be replicated via S3 replication to a US-WEST location. They can also be regenerated by the application though this takes days to do if all the files are lost. Replication requires versioning, so this is enabled but unlikely to be used.

Minimal public recovery requires the following data:

  • Postgres database

  • Original binaries for downloading tiffs

  • Derivative files

If we have an AWS outage affecting our region, the fastest recovery options are to (if needed) rebuild the servers in another region and edit the local_env.yml file in either ansible or by hand to point to the backup S3 buckets for the original and derivative files. The postgres database will need to be downloaded from S3 and installed onto the new servers if we needed to use new servers. After that point all current public facing aspects will be restored. Since the backup buckets do not sync with the original data buckets, staff should not upload new files though they can edit metadata on existing works. (Once the original S3 buckets have had service restored or their data copied back, set the application to use the original buckets with local_env.yml and users can now add items. The postgres database may need to be copied back to the original server(s) or region if a new server setup was used.

In the case of smaller issues, like single file corruption or deletion, the simplest method for original files is to locate them on S3 and look at previous versions. We keep 30 days worth of versions so if the error was found within a month you should be able to revert back to an earlier file. For derivatives it is easier to simply regenerate them via the command line.

  • No labels