Cassandra backup and recovery

This section discusses how to configure data backup and recovery for the Apache Cassandra database ring installed in the Apigee hybrid runtime plane. See also Cassandra database.

What you need to know about Cassandra backups

Cassandra is a replicated database that is configured to have at least 3 copies of your data in each region or data center. Cassandra uses streaming replication and read repairs to maintain the data replicas in each region or data center at any given point.

In hybrid, Cassandra backups are not enabled by default. It's a good practice, however, to enable Cassandra backups in case your data is accidentally deleted.

What is backed up?

The backup configuration described in this topic backs up the following entities:

  • Cassandra schema including the user schema (Apigee keyspace definitions)
  • Cassandra partition token information per node
  • A snapshot of the Cassandra data

Where is backup data stored?

Backed up data is stored in a Google Cloud Storage (GCS) bucket that you must create. Bucket creation and configuration is covered in this topic.

Scheduling Cassandra backups

Backups are scheduled as cron jobs in the runtime plane. To schedule Cassandra backups:

  1. Run the following create-service-account command to create a GCP service account (SA) with the standard roles/storage.objectAdmin role. This SA role allows you to write backup data to Google Cloud Storage (GCS). Execute the following command in the hybrid installation root directory: 
    ./tools/create-service-account apigee-cassandra output-dir
     For example: 
    ./tools/create-service-account apigee-cassandra ./service-accounts
     For more information about GCP service accounts, see Creating and managing service accounts.
  2. The create-service-account command saves a JSON file containing the service account private key. The file is saved in the same directory where the command executes. You will need the path to this file in the following steps.
  3. Create a GCS bucket. Specify a reasonable data retention policy for the bucket. Apigee recommends a data retention policy of 15 days.
  4. Open your overrides.yaml file.
  5. Add the following cassandra.backup properties to enable backup. Do not remove any of the properties that are already configured. 
    cassandra:
  ...

  backup:
    enabled: true
    serviceAccountPath: sa_json_file_path
    dbStorageBucket: gcs_bucket_path
    schedule: backup_schedule_code

  ...
     Where:
    Property Description
    enabled Backup is disabled by default. You must set this property to true
    serviceAccountPath The path on your filesystem to the service account JSON file that was downloaded when you ran ./tools/create-service-account
    dbStorageBucket GCS storage bucket path in this format: gs://bucket_name. The gs:// is required.
    schedule The time when the backup starts, specified in standard crontab syntax. Default: 0 2 * * *

    Note: Avoid scheduling a backup that starts a short time after you apply the backup configuration to your cluster. When you apply the backup configuration, Kubernetes recreates the Cassandra nodes. If the backup starts before the nodes restart (possibly several minutes) the backup will fail.

    For example: 
    ...

cassandra:
  storage:
    type: gcepd
    capacity: 50Gi
    gcepd:
      replicationType: regional-pd
  sslRootCAPath: "/Users/myhome/ssh/cassandra.crt"
  sslCertPath: "/Users/myhome/ssh/cassandra.crt"
  sslKeyPath: "/Users/myhome/ssh/cassandra.key"
  auth:
    default:
      password: "abc123"
    admin:
      password: "abc234"
    ddl:
      password: "abc345"
    dml:
      password: "abc456"
  nodeSelector:
    key: cloud.google.com/gke-nodepool
    value: apigee-data
  backup:
    enabled: true
    serviceAccountPath: "/Users/myhome/.ssh/my_cassandra_backup.json"
    dbStorageBucket: "gs://myname-cassandra-backup"
    schedule: "45 23 * * 6"

  ...
  6. Apply the configuration changes to the new cluster. For example: 
    ./apigeectl apply -c cassandra -f my-overrides.yaml

Restoring backups

Restoration takes the data from the backup location and restores it into a new Cassandra cluster with the same number of pods. The new cluster must have a namespace that is different than your runtime plane cluster.

To restore Cassandra backups:

  1. Create a new Kubernetes cluster with a new namespace. You cannot use the same cluster/namespace that you used for the original hybrid installation.
  2. In the root hybrid installation directory, create a new overrides-restore.yaml file.
  3. Copy the complete Cassandra configuration from your original overrides.yaml file into the new one.
  4. Add a namespace element. Do not use the same namespace you used for your original cluster.
    5. namespace: your-restore-namespace

cassandra:
  storage:
    type: gcepd
    capacity: 50Gi
    gcepd:
      replicationType: regional-pd
  nodeSelector:
    key: cloud.google.com/gke-nodepool
    value: apigee-data
  sslRootCAPath: path_to_root_ca_file
  sslCertPath: path_to_ssl_cert_file
  sslKeyPath: path_to_ssl_key_file
  auth:
    default:
      password: your_cassandra_password
    admin:
      password: admin_password
    ddl:
      password: ddl_password
    dml:
      password: dml_password

  restore:
      enabled: true
      snapshotTimestamp: timestamp
      serviceAccountPath: sa_json_file_path
      dbStorageBucket: gcs_bucket_path
      image:
        pullPolicy: Always
     Where:
    Property Description
    ssl*Path, auth.* Use the same TLS auth credentials you used to create the original Cassandra database.
    snapshotTimestamp The timestamp of the backup snapshot to restore.
    serviceAccountPath The path on your filesystem to the service account you created for the backup.
    dbStorageBucket GCS storage bucket path where your backup is stored, in this format: gs://bucket_name. The gs:// is required.
    For example: 
    namespace: cassandra-restore

cassandra:
  storage:
    type: gcepd
    capacity: 50Gi
    gcepd:
      replicationType: regional-pd
  sslRootCAPath: "/Users/myhome/ssh/cassandra.crt"
  sslCertPath: "/Users/myhome/ssh/cassandra.crt"
  sslKeyPath: "/Users/myhome/ssh/cassandra.key"
  auth:
    default:
      password: "abc123"
    admin:
      password: "abc234"
    ddl:
      password: "abc345"
    dml:
      password: "abc456"
  nodeSelector:
    key: cloud.google.com/gke-nodepool
    value: apigee-data
  restore:
    enabled: true
    snapshotTimestamp: "20190417002207"
    serviceAccountPath: "/Users/myhome/.ssh/my_cassandra_backup.json"
    dbStorageBucket: "gs://myname-cassandra-backup"
    image:
      pullPolicy: Always

    Where snapshotTimestamp is the timestamp associated with the backup you are restoring.

  5. Create the new Cassandra cluster: 
      ./apigeectl apply -c cassandra -f ./overrides-restore.yaml

Viewing the restore logs

You can check the restore job logs and grep for error to make sure the restore log has no errors.

Verify the restore completed

To check if the restore operation completed: 

kubectl get pods

NAME                           READY     STATUS      RESTARTS   AGE
apigee-cassandra-0             1/1       Running     0          1h
apigee-cassandra-1             1/1       Running     0          1h
apigee-cassandra-2             1/1       Running     0          59m
apigee-cassandra-restore-b4lgf 0/1       Completed   0          51m

View the restore logs

To view the restore logs: 

kubectl logs -f apigee-cassandra-restore-b4lgf

Restore Logs:

Activated service account credentials for: [apigee-cassandra-backup-svc@gce-myusername.iam.gserviceaccount.com]
to download file gs://gce-myusername-apigee-cassandra-backup/apigeecluster/dc-1/backup_20190405011309_schema.tgz
INFO: download sucessfully extracted the backup files from gs://gce-myusername-apigee-cassandra-backup/apigeecluster/dc-1
finished downloading schema.cql
to create schema from 10.32.0.28

Warnings :
dclocal_read_repair_chance table option has been deprecated and will be removed in version 4.0

dclocal_read_repair_chance table option has been deprecated and will be removed in version 4.0


Warnings :
dclocal_read_repair_chance table option has been deprecated and will be removed in version 4.0

dclocal_read_repair_chance table option has been deprecated and will be removed in version 4.0

INFO: the schema has been restored
starting apigee-cassandra-0 in default
starting apigee-cassandra-1 in default
starting apigee-cassandra-2 in default
84 95 106
waiting on waiting nodes $pid to finish  84
Activated service account credentials for: [apigee-cassandra-backup-svc@gce-myusername.iam.gserviceaccount.com]
Activated service account credentials for: [apigee-cassandra-backup-svc@gce-myusername.iam.gserviceaccount.com]
Activated service account credentials for: [apigee-cassandra-backup-svc@gce-myusername.iam.gserviceaccount.com]
INFO: restore downloaded  tarball and extracted the file from  gs://gce-myusername-apigee-cassandra-backup/apigeecluster/dc-1
INFO: restore downloaded  tarball and extracted the file from  gs://gce-myusername-apigee-cassandra-backup/apigeecluster/dc-1
INFO: restore downloaded  tarball and extracted the file from  gs://gce-myusername-apigee-cassandra-backup/apigeecluster/dc-1
INFO  12:02:28 Configuration location: file:/etc/cassandra/cassandra.yaml
…...

INFO  12:02:41 [Stream #e013ee80-5863-11e9-8458-353e9e3cb7f9] All sessions completed

Summary statistics:
   Connections per host    : 3
   Total files transferred : 2
   Total bytes transferred : 0.378KiB
   Total duration          : 5048 ms
   Average transfer rate   : 0.074KiB/s
   Peak transfer rate      : 0.075KiB/s

progress: [/10.32.1.155]0:1/1 100% 1:1/1 100% [/10.32.0.28]1:1/1 100% 0:1/1 100% [/10.32.3.220]0:1/1 100% 1:1/1 100% total: 100% 0.000KiB/s (avg: 0.074KiB/s)
INFO  12:02:41 [Stream #e013ee80-5863-11e9-8458-353e9e3cb7f9] All sessions completed
progress: [/10.32.1.155]0:1/1 100% 1:1/1 100% [/10.32.0.28]1:1/1 100% 0:1/1 100% [/10.32.3.220]0:1/1 100% 1:1/1 100% total: 100% 0.000KiB/s (avg: 0.074KiB/s)
INFO  12:02:41 [Stream #e013ee80-5863-11e9-8458-353e9e3cb7f9] All sessions completed
INFO  12:02:41 [Stream #e013ee80-5863-11e9-8458-353e9e3cb7f9] All sessions completed
INFO: ./apigee/data/cassandra/data/ks1/user-9fbae960571411e99652c7b15b2db6cc restored successfully
INFO: Restore 20190405011309 completed
INFO: ./apigee/data/cassandra/data/ks1/user-9fbae960571411e99652c7b15b2db6cc restored successfully
INFO: Restore 20190405011309 completed
waiting on waiting nodes $pid to finish  106
Restore finished

Verify backup job

You can also verify your backup job after your backup cronjob is scheduled. After the cronjob has been scheduled, you should see something like this: 

kubectl get pods
NAME                        READY     STATUS      RESTARTS   AGE
apigee-cassandra-0          1/1       Running     0          2h
apigee-cassandra-1          1/1       Running     0          2h
apigee-cassandra-2          1/1       Running     0          2h
apigee-cassandra-backup-1554515580-pff6s   0/1       Running     0          54s

Check the backup logs

The backup job:

  • Creates a schema.cql file.
  • Uploads it to your storage bucket.
  • Echoes the node to backup the data and uploads it at the same time.
  • Waits until all of the data is uploaded.
kubectl logs -f apigee-cassandra-backup-1554515580-pff6s

myusername-macbookpro:cassandra-backup-utility myusername$ kubectl logs -f apigee-cassandra-backup-1554577680-f9sc4
starting apigee-cassandra-0 in default
starting apigee-cassandra-1 in default
starting apigee-cassandra-2 in default
35 46 57
waiting on process  35
Activated service account credentials for: [apigee-cassandra-backup-svc@gce-myusername.iam.gserviceaccount.com]
Activated service account credentials for: [apigee-cassandra-backup-svc@gce-myusername.iam.gserviceaccount.com]
Activated service account credentials for: [apigee-cassandra-backup-svc@gce-myusername.iam.gserviceaccount.com]
Requested creating snapshot(s) for [all keyspaces] with snapshot name [20190406190808] and options {skipFlush=false}
Snapshot directory: 20190406190808
INFO: backup created cassandra snapshot 20190406190808
tar: Removing leading `/' from member names
/apigee/data/cassandra/data/ks1/mytest3-37bc2df0587811e98e8d875b0ed64754/snapshots/
/apigee/data/cassandra/data/ks1/mytest3-37bc2df0587811e98e8d875b0ed64754/snapshots/20190406190808/
/apigee/data/cassandra/data/ks1/mytest3-37bc2df0587811e98e8d875b0ed64754/snapshots/20190406190808/mc-1-big-Data.db
Requested creating snapshot(s) for [all keyspaces] with snapshot name [20190406190808] and options {skipFlush=false}
Requested creating snapshot(s) for [all keyspaces] with snapshot name [20190406190808] and options {skipFlush=false}
Snapshot directory: 20190406190808
INFO: backup created cassandra snapshot 20190406190808
tar: Removing leading `/' from member names
/apigee/data/cassandra/data/system/hints-2666e20573ef38b390fefecf96e8f0c7/snapshots/
/apigee/data/cassandra/data/system/hints-2666e20573ef38b390fefecf96e8f0c7/snapshots/20190406190808/
/apigee/data/cassandra/data/system/hints-2666e20573ef38b390fefecf96e8f0c7/snapshots/20190406190808/manifest.json
/apigee/data/cassandra/data/system/prepared_statements-18a9c2576a0c3841ba718cd529849fef/snapshots/
/apigee/data/cassandra/data/system/prepared_statements-18a9c2576a0c3841ba718cd529849fef/snapshots/20190406190808/
/apigee/data/cassandra/data/system/prepared_statements-18a9c2576a0c3841ba718cd529849fef/snapshots/20190406190808/manifest.json
/apigee/data/cassandra/data/system/range_xfers-55d764384e553f8b9f6e676d4af3976d/snapshots/
/apigee/data/cassandra/data/system/range_xfers-55d764384e553f8b9f6e676d4af3976d/snapshots/20190406190808/
/apigee/data/cassandra/data/system/range_xfers-55d764384e553f8b9f6e676d4af3976d/snapshots/20190406190808/manifest.json
/apigee/data/cassandra/data/system/peer_events-59dfeaea8db2334191ef109974d81484/snapshots/
/apigee/data/cassandra/data/system/peer_events-59dfeaea8db2334191ef109974d81484/snapshots/20190406190808/
/apigee/data/cassandra/data/system/peer_events-59dfeaea8db2334191ef109974d81484/snapshots/20190406190808/manifest.json
/apigee/data/cassandra/data/system/built_views-4b3c50a9ea873d7691016dbc9c38494a/snapshots/
/apigee/data/cassandra/data/system/built_views-4b3c50a9ea873d7691016dbc9c38494a/snapshots/20190406190808/
/apigee/data/cassandra/data/system/built_views-4b3c50a9ea873d7691016dbc9c38494a/snapshots/20190406190808/manifest.json
……
/apigee/data/cassandra/data/ks2/user-d6d39d70586311e98e8d875b0ed64754/snapshots/
/apigee/data/cassandra/data/ks2/user-d6d39d70586311e98e8d875b0ed64754/snapshots/20190406190808/
/apigee/data/cassandra/data/ks2/user-d6d39d70586311e98e8d875b0ed64754/snapshots/20190406190808/mc-3-big-Filter.db
/apigee/data/cassandra/data/ks2/user-d6d39d70586311e98e8d875b0ed64754/snapshots/20190406190808/mc-2-big-CompressionInfo.db
/apigee/data/cassandra/data/ks2/user-d6d39d70586311e98e8d875b0ed64754/snapshots/20190406190808/mc-2-big-Index.db
/apigee/data/cassandra/data/ks2/user-d6d39d70586311e98e8d875b0ed64754/snapshots/20190406190808/mc-3-big-Statistics.db
/apigee/data/cassandra/data/ks2/user-d6d39d70586311e98e8d875b0ed64754/snapshots/20190406190808/mc-2-big-Data.db
/apigee/data/cassandra/data/ks2/user-d6d39d70586311e98e8d875b0ed64754/snapshots/20190406190808/mc-1-big-Index.db
/apigee/data/cassandra/data/ks2/user-d6d39d70586311e98e8d875b0ed64754/snapshots/20190406190808/mc-1-big-Statistics.db
/apigee/data/cassandra/data/ks2/user-d6d39d70586311e98e8d875b0ed64754/snapshots/20190406190808/mc-2-big-TOC.txt
/apigee/data/cassandra/data/ks2/user-d6d39d70586311e98e8d875b0ed64754/snapshots/20190406190808/mc-2-big-Statistics.db
/apigee/data/cassandra/data/ks2/user-d6d39d70586311e98e8d875b0ed64754/snapshots/20190406190808/mc-1-big-Summary.db
/apigee/data/cassandra/data/ks2/user-d6d39d70586311e98e8d875b0ed64754/snapshots/20190406190808/mc-1-big-Filter.db
/apigee/data/cassandra/data/ks2/user-d6d39d70586311e98e8d875b0ed64754/snapshots/20190406190808/mc-2-big-Summary.db
/apigee/data/cassandra/data/ks2/user-d6d39d70586311e98e8d875b0ed64754/snapshots/20190406190808/mc-3-big-Index.db
/apigee/data/cassandra/data/ks2/user-d6d39d70586311e98e8d875b0ed64754/snapshots/20190406190808/manifest.json
/apigee/data/cassandra/data/ks2/user-d6d39d70586311e98e8d875b0ed64754/snapshots/20190406190808/mc-2-big-Filter.db
/apigee/data/cassandra/data/ks2/user-d6d39d70586311e98e8d875b0ed64754/snapshots/20190406190808/mc-2-big-Digest.crc32
/apigee/data/cassandra/data/ks2/user-d6d39d70586311e98e8d875b0ed64754/snapshots/20190406190808/mc-3-big-Summary.db
/apigee/data/cassandra/data/ks2/user-d6d39d70586311e98e8d875b0ed64754/snapshots/20190406190808/mc-3-big-Data.db
/apigee/data/cassandra/data/ks2/user-d6d39d70586311e98e8d875b0ed64754/snapshots/20190406190808/mc-1-big-TOC.txt
/apigee/data/cassandra/data/ks2/user-d6d39d70586311e98e8d875b0ed64754/snapshots/20190406190808/schema.cql
/apigee/data/cassandra/data/ks2/user-d6d39d70586311e98e8d875b0ed64754/snapshots/20190406190808/mc-3-big-CompressionInfo.db
/apigee/data/cassandra/data/ks2/user-d6d39d70586311e98e8d875b0ed64754/snapshots/20190406190808/mc-1-big-Digest.crc32
/apigee/data/cassandra/data/ks2/user-d6d39d70586311e98e8d875b0ed64754/snapshots/20190406190808/mc-3-big-TOC.txt
/apigee/data/cassandra/data/ks2/user-d6d39d70586311e98e8d875b0ed64754/snapshots/20190406190808/mc-1-big-Data.db
/apigee/data/cassandra/data/ks2/user-d6d39d70586311e98e8d875b0ed64754/snapshots/20190406190808/mc-3-big-Digest.crc32
/apigee/data/cassandra/data/ks2/user-d6d39d70586311e98e8d875b0ed64754/snapshots/20190406190808/mc-1-big-CompressionInfo.db
……
/tmp/tokens.txt
/ [1 files][    0.0 B/    0.0 B]
Operation completed over 1 objects.
/ [1 files][    0.0 B/    0.0 B]
Operation completed over 1 objects.
INFO: backup created tarball and transfered the file to gs://gce-myusername-apigee-cassandra-backup/apigeecluster/dc-1
INFO: removing cassandra snapshot
INFO: backup created tarball and transfered the file to gs://gce-myusername-apigee-cassandra-backup/apigeecluster/dc-1
INFO: removing cassandra snapshot
Requested clearing snapshot(s) for [all keyspaces]
INFO: Backup 20190406190808 completed
waiting on process  46
Requested clearing snapshot(s) for [all keyspaces]
INFO: Backup 20190406190808 completed
Requested clearing snapshot(s) for [all keyspaces]
waiting on process  57
INFO: Backup 20190406190808 completed
waiting result
to get schema from 10.32.0.28
INFO: /tmp/schema.cql has been generated
Activated service account credentials for: [apigee-cassandra-backup-svc@gce-myusername.iam.gserviceaccount.com]
tar: removing leading '/' from member names
tmp/schema.cql
Copying from ...
/ [1 files][    0.0 B/    0.0 B]
Operation completed over 1 objects.
INFO: backup created tarball and transfered the file to gs://gce-myusername-apigee-cassandra-backup/apigeecluster/dc-1
finished uploading schema.cql