Backing Up Data

AttentionThis page documents an earlier version. Go to the latest (v2.1)version.

Overview

This page covers backups for YugabyteDB CE. Here are some points to keep in mind.

• Export-based backups (YugabyteDB CE)

  • Single row ACID backups
  • Backup the schema and data separately
  • Multi-threaded parallelism

• Distributed backups (YugabyteDB EE)

  • Single tablet ACID backups
  • Integrated backup solution, integrated with object stores such as AWS S3
  • Massively parallel, efficient for very large data sets

Dump the keyspace schema (optional) as well as the data in the tables to create a backup of the data in YugabyteDB. Here are some points to bear in mind.

Schema backup

• Backing up schema for one keyspace

In order to backup the schema for a particular keyspace, run the following command.

$ cqlsh -e "DESC KEYSPACE <keyspace name>" > schema.cql

• Backing up schema for entire cluster

In order to backup the schema for all tables across all keyspaces, run the following command.

cqlsh -e "DESC SCHEMA" > schema.cql

Data backup

The following command exports all the data from a table in the CSV (comma separated value) format to the output file specified. Each row in the table is written to a separate line in the file with the various column values separated by the delimiter.

• Backing up all columns of the table

All columns of the table are exported by default.

cqlsh -e "COPY <keyspace>.<table> TO 'data.csv' WITH HEADER = TRUE;"

• Backing up select columns of the table

In order to backup selected columns of the table, specify the column names in a list.

cqlsh -e "COPY <keyspace>.<table> (<column 1 name>, <column 2 name>, ...) TO 'data.csv' WITH HEADER = TRUE;"

Options

• Connecting to a remote host and port

The default host is 127.0.0.1 and the default port is 9042. You can override these values as shown below.

cqlsh -e <command> <host> [<port>]

• Copy options

The syntax to specify options in the COPY TO command is shown below.

COPY table_name [( column_list )]
FROM 'file_name'[, 'file2_name', ...] | STDIN
[WITH option = 'value' [AND ...]]

There are a number of useful options in the COPY TO command used to perform the backup. Some of these are outlined below.

Example

We are going to use the example shown in the quick start section in order to demonstrate how to perform backups.

This section assumes you already have a YugabyteDB cluster. You can install a local cluster on your laptop using these quick start instructions.

Create a table with data

Create a keyspace for the stock ticker app.

cqlsh> CREATE KEYSPACE stocks;

Create the stock ticker table.

cqlsh> CREATE TABLE myapp.stock_market (
    stock_symbol text,
    ts text,
    current_price float,
    PRIMARY KEY (stock_symbol, ts)
);

Insert some sample data.

INSERT INTO myapp.stock_market (stock_symbol,ts,current_price) VALUES ('AAPL','2017-10-26 09:00:00',157.41);
INSERT INTO myapp.stock_market (stock_symbol,ts,current_price) VALUES ('AAPL','2017-10-26 10:00:00',157);
INSERT INTO myapp.stock_market (stock_symbol,ts,current_price) VALUES ('FB','2017-10-26 09:00:00',170.63);
INSERT INTO myapp.stock_market (stock_symbol,ts,current_price) VALUES ('FB','2017-10-26 10:00:00',170.1);
INSERT INTO myapp.stock_market (stock_symbol,ts,current_price) VALUES ('GOOG','2017-10-26 09:00:00',972.56);
INSERT INTO myapp.stock_market (stock_symbol,ts,current_price) VALUES ('GOOG','2017-10-26 10:00:00',971.91);

You can query all the 6 rows we inserted by running the following command in cqlsh.

cqlsh> SELECT * FROM myapp.stock_market;

 stock_symbol | ts                  | current_price
--------------+---------------------+---------------
         GOOG | 2017-10-26 09:00:00 |        972.56
         GOOG | 2017-10-26 10:00:00 |     971.90997
         AAPL | 2017-10-26 09:00:00 |        157.41
         AAPL | 2017-10-26 10:00:00 |           157
           FB | 2017-10-26 09:00:00 |        170.63
           FB | 2017-10-26 10:00:00 |     170.10001
(6 rows)

Backup the schema

Run the following in order to backup the schema of the keyspace myapp.

cqlsh -e "DESC KEYSPACE myapp" > myapp_schema.cql

The schema of the keyspace myapp along with the tables in it are saved to the file myapp_schema.cql.

$ cat myapp_schema.cql

CREATE KEYSPACE myapp WITH replication = {'class': 'SimpleStrategy', 'replication_factor': '3'}  AND durable_writes = true;
CREATE TABLE myapp.stock_market (
    stock_symbol text,
    ts text,
    current_price float,
    PRIMARY KEY (stock_symbol, ts)
) WITH CLUSTERING ORDER BY (ts ASC)
    AND default_time_to_live = 0;

Backing up all the columns of the table

Run the following command in order to backup the data in the table myapp.stock_market.

$ cqlsh -e "COPY myapp.stock_market TO 'myapp_data.csv' WITH HEADER = TRUE ;"

All columns of the rows in the table myapp.stock_market are saved to the file myapp_data.csv.

$ cat myapp_data.csv
stock_symbol,ts,current_price
AAPL,2017-10-26 09:00:00,157.41
AAPL,2017-10-26 10:00:00,157
FB,2017-10-26 09:00:00,170.63
FB,2017-10-26 10:00:00,170.10001
GOOG,2017-10-26 09:00:00,972.56
GOOG,2017-10-26 10:00:00,971.90997

Backing up some columns of the table

In order to backup a subset of columns, you can specify them in the backup command. In the example below, the stock_symbol and ts columns are backed up, while the current_price column is not.

$ cqlsh -e "COPY myapp.stock_market (stock_symbol, ts) TO 'myapp_data_partial.csv' WITH HEADER = TRUE ;"

The selected columns (stock_symbol and ts) of the rows in the table myapp.stock_market are saved to the file myapp_data_partial.csv.

$ cat myapp_data_partial.csv
stock_symbol,ts
AAPL,2017-10-26 09:00:00
AAPL,2017-10-26 10:00:00
FB,2017-10-26 09:00:00
FB,2017-10-26 10:00:00
GOOG,2017-10-26 09:00:00
GOOG,2017-10-26 10:00:00