Was this page helpful?
The Scylla Manager backup requires a backup location string that specifies the storage provider and name of a bucket (using Amazon S3 naming) ex. s3:<your S3 bucket name>.
In that bucket Scylla Manager creates a buckup directory where all the backup data and metadata are stored.
There are three subdirectories:
meta - contains backup manifests
schema - contains text dumps of database schema
sst - contains the data files
The meta directory contains backup manifests. Please find below a listing of meta directory with three nodes and three runs of a backup task.
/backup
└── meta
└── cluster
└── 3e99d4a8-67d2-45fe-87fb-87b1b90ea2dc
└── dc
└── dc1
└── node
├── 01c9349e-89e6-4ceb-a727-4f27f9f2acce
│ ├── task_f70117d8-c10e-4e90-9606-2587936b3757_tag_sm_20210809095541UTC_manifest.json.gz
│ ├── task_f70117d8-c10e-4e90-9606-2587936b3757_tag_sm_20210809095542UTC_manifest.json.gz
│ └── task_f70117d8-c10e-4e90-9606-2587936b3757_tag_sm_20210809095748UTC_manifest.json.gz
├── 427be6b6-0773-465c-b1a6-4ed2265500fe
│ ├── task_f70117d8-c10e-4e90-9606-2587936b3757_tag_sm_20210809095541UTC_manifest.json.gz
│ ├── task_f70117d8-c10e-4e90-9606-2587936b3757_tag_sm_20210809095542UTC_manifest.json.gz
│ └── task_f70117d8-c10e-4e90-9606-2587936b3757_tag_sm_20210809095748UTC_manifest.json.gz
└── ae6a5cf3-cb53-4954-8c16-866003727111
├── task_f70117d8-c10e-4e90-9606-2587936b3757_tag_sm_20210809095541UTC_manifest.json.gz
├── task_f70117d8-c10e-4e90-9606-2587936b3757_tag_sm_20210809095542UTC_manifest.json.gz
└── task_f70117d8-c10e-4e90-9606-2587936b3757_tag_sm_20210809095748UTC_manifest.json.gz
Each node has it’s own directory, the path is structured as follows.
/backup/meta/cluster/<cluster ID>/dc/<data center name>/node/<node ID>
The directory is flat and contains all the manifests, the file name is structured as follows.
task_<task ID>_tag_<snapshot tag>_manifest.json.gz
You may find that the directory contains manifest files with .tmp suffix.
Those are manifests files of backups that are being uploaded to the backup location.
They are are marked as temporary until all the backup files are fully uploaded.
The schema directory contains text dumps of database schema.
Starting from ScyllaDB 6.0 and 2024.2 (and compatible ScyllaDB Manager 3.3), they are required for the schema restoration.
They have schema_with_internals.json.gz suffix and represent the output of DESCRIBE SCHEMA WITH INTERNALS CQL query,
which returns rows in the format keyspace|type|name|CQL create statement in the correct order.
If you are using an earlier ScyllaDB version, those files are mainly for information purposes.
To restore the backup, use the system_schema keyspace.
They have schema.tar.gz suffix and represent schema archive divided among keyspaces.
To enable upload of the files make sure that the cluster is added with username and password flags.
/backup
└── schema
└── cluster
└── 3e99d4a8-67d2-45fe-87fb-87b1b90ea2dc
├── task_f70117d8-c10e-4e90-9606-2587936b3757_tag_sm_20210809095541UTC_schema_with_internals.json.gz
├── task_f70117d8-c10e-4e90-9606-2587936b3757_tag_sm_20210809095542UTC_schema_with_internals.json.gz
└── task_f70117d8-c10e-4e90-9606-2587936b3757_tag_sm_20210809095748UTC_schema_with_internals.json.gz
The schema file path is structured as follows.
/backup/schema/cluster/<cluster ID>/task_<task ID>_tag_<snapshot tag>_schema_with_internals.json.gz
The sst directory contains the data files. Please find below a part listing of sst directory showing a single table of a single node.
/backup
└── sst/
└── cluster
└── 3e99d4a8-67d2-45fe-87fb-87b1b90ea2dc
└── dc
└── dc1
└── node
├── 01c9349e-89e6-4ceb-a727-4f27f9f2acce
│ └── keyspace
│ └── backuptest_data
│ └── table
│ └── big_table
│ └── f34b6ff0f8f711eb9fcf000000000000
│ ├── mc-3ggs_0xmx_3261s2qpoyoxpg4min-big-CompressionInfo.db
│ ├── mc-3ggs_0xmx_3261s2qpoyoxpg4min-big-Data.db
│ ├── mc-3ggs_0xmx_3261s2qpoyoxpg4min-big-Digest.crc32
│ ├── mc-3ggs_0xmx_3261s2qpoyoxpg4min-big-Filter.db
│ ├── mc-3ggs_0xmx_3261s2qpoyoxpg4min-big-Index.db
│ ├── mc-3ggs_0xmx_3261s2qpoyoxpg4min-big-Scylla.db
│ ├── mc-3ggs_0xmx_3261s2qpoyoxpg4min-big-Statistics.db
│ ├── mc-3ggs_0xmx_3261s2qpoyoxpg4min-big-Summary.db
│ ├── mc-3ggs_0xmx_3261s2qpoyoxpg4min-big-TOC.txt
│ ├── mc-3ggs_0xmx_3nlnl24aeuqnqzxgxr-big-CompressionInfo.db
│ ├── mc-3ggs_0xmx_3nlnl24aeuqnqzxgxr-big-Data.db
│ ├── mc-3ggs_0xmx_3nlnl24aeuqnqzxgxr-big-Digest.crc32
│ ├── mc-3ggs_0xmx_3nlnl24aeuqnqzxgxr-big-Filter.db
│ ├── mc-3ggs_0xmx_3nlnl24aeuqnqzxgxr-big-Index.db
│ ├── mc-3ggs_0xmx_3nlnl24aeuqnqzxgxr-big-Scylla.db
│ ├── mc-3ggs_0xmx_3nlnl24aeuqnqzxgxr-big-Statistics.db
│ ├── mc-3ggs_0xmx_3nlnl24aeuqnqzxgxr-big-Summary.db
│ ├── mc-3ggs_0xmx_3nlnl24aeuqnqzxgxr-big-TOC.txt
│ ├── mc-3ggs_0xlz_1u89d24aeuqnqzxgxr-big-CompressionInfo.db
│ ├── mc-3ggs_0xlz_1u89d24aeuqnqzxgxr-big-Data.db
│ ├── mc-3ggs_0xlz_1u89d24aeuqnqzxgxr-big-Digest.crc32
│ ├── mc-3ggs_0xlz_1u89d24aeuqnqzxgxr-big-Filter.db
│ ├── mc-3ggs_0xlz_1u89d24aeuqnqzxgxr-big-Index.db
│ ├── mc-3ggs_0xlz_1u89d24aeuqnqzxgxr-big-Scylla.db
│ ├── mc-3ggs_0xlz_1u89d24aeuqnqzxgxr-big-Statistics.db
│ ├── mc-3ggs_0xlz_1u89d24aeuqnqzxgxr-big-Summary.db
│ └── mc-3ggs_0xlz_1u89d24aeuqnqzxgxr-big-TOC.txt
Each node has it’s own directory, the path is structured as follows.
/backup/sst/cluster/<cluster ID>/dc/<data center name>/node/<node ID>
Under the node directory each table version has it’s directory, the path is structured as follows.
keyspace/<keyspace name>/table/<table name>/<table schema version>
The directory contains all the table files. Some files may be used in more than one backup.
You may also find that some of those files have the snapshot tag suffix (e.g. .sm_20210809095541UTC).
Even though sstables (data files) are immutable by nature, using non UUID sstable identifiers alongside
replacing a running node
makes it possible to have two different sstables originating
from the same table and with the same node ID. In this case, older versions of sstable are appended with the snapshot tag
indicating the backup that introduced newer version of given sstable. The most current version of sstable does not have the snapshot tag suffix.
For example, sstable mc-5-big-Data.db will be used when restoring snapshot sm_20210809095541UTC,
while mc-5-big-Data.db.sm_20210809095541UTC will be used when restoring sm_20210809095542UTC or sm_20210809095748UTC snapshots
(lexicographical order on snapshot tags is also the chronological order).
Scylla Manager Manifest files are gzipped JSON files. Each node has it’s own manifest file. If a cluster has three nodes a backup would contain three manifest files with the same name but under different directories. Please find below the contents of the manifest file of the node shown in the sst section.
{
"version": "v2",
"cluster_name": "test_cluster",
"ip": "192.168.100.13",
"index": [
{
"keyspace": "backuptest_data",
"table": "big_table",
"version": "f34b6ff0f8f711eb9fcf000000000000",
"files": [
"mc-3ggs_0xmx_3261s2qpoyoxpg4min-big-CompressionInfo.db",
"mc-3ggs_0xmx_3261s2qpoyoxpg4min-big-Data.db",
"mc-3ggs_0xmx_3261s2qpoyoxpg4min-big-Digest.crc32",
"mc-3ggs_0xmx_3261s2qpoyoxpg4min-big-Filter.db",
"mc-3ggs_0xmx_3261s2qpoyoxpg4min-big-Index.db",
"mc-3ggs_0xmx_3261s2qpoyoxpg4min-big-Scylla.db",
"mc-3ggs_0xmx_3261s2qpoyoxpg4min-big-Statistics.db",
"mc-3ggs_0xmx_3261s2qpoyoxpg4min-big-Summary.db",
"mc-3ggs_0xmx_3261s2qpoyoxpg4min-big-TOC.txt",
"mc-3ggs_0xmx_3nlnl24aeuqnqzxgxr-big-CompressionInfo.db",
"mc-3ggs_0xmx_3nlnl24aeuqnqzxgxr-big-Data.db",
"mc-3ggs_0xmx_3nlnl24aeuqnqzxgxr-big-Digest.crc32",
"mc-3ggs_0xmx_3nlnl24aeuqnqzxgxr-big-Filter.db",
"mc-3ggs_0xmx_3nlnl24aeuqnqzxgxr-big-Index.db",
"mc-3ggs_0xmx_3nlnl24aeuqnqzxgxr-big-Scylla.db",
"mc-3ggs_0xmx_3nlnl24aeuqnqzxgxr-big-Statistics.db",
"mc-3ggs_0xmx_3nlnl24aeuqnqzxgxr-big-Summary.db",
"mc-3ggs_0xmx_3nlnl24aeuqnqzxgxr-big-TOC.txt"
],
"size": 1256031
}
],
"size": 1256031,
"tokens": [
-9214072223864974000,
-9209525598183111000,
-9203154907091372000,
-9121005554342506000,
-9077102529676286000,
-8972056514211153000,
-8928968163169332000,
-8928882009746142000,
-8863551618551160000,
-8857300148094569000,
...],
"schema": "backup/schema/cluster/3e99d4a8-67d2-45fe-87fb-87b1b90ea2dc/task_f70117d8-c10e-4e90-9606-2587936b3757_tag_sm_20210809095541UTC_schema.tar.gz"
}
The manifest contains the following information.
version - the version of the manifest
cluster_name - name of the cluster as registered in Scylla Manager
ip - public IP address of the node
index - list of tables, each table holds a list of file names
size - total size of files in index
tokens - tokens owned by node, they allow to recreate the cluster topology
schema - path to schema file
Was this page helpful?