Systems and methods for backup and restore of master-less distributed database clusters

1. A machine implemented method for a distributed, master-less database cluster having a plurality of nodes each executing a database application and storing at least a portion of a database for the cluster, comprising:

capturing a topology of the distributed, master-less database cluster prior to a first phase of a backup operation;

associating token information of each node with a logical storage object of a shared storage system that stores data of the database for each node on behalf of the database application; wherein the token information is based on a hash value of a portion of the data of the database;

generating a snapshot of all logical storage objects associated with each of the plurality of nodes during the first phase;

storing token information of each node with backup metadata for the snapshot of all the logical storage objects during the first phase, where the token information determines what data is stored by each node for the database;

initiating a second phase of the backup operation at a recovery node using the backup metadata from the first phase of the backup operation;

cloning snapshots of all the logical storage objects from the first phase of the backup operation;

flushing information from a commit log for each node to data containers of the database;

creating a new backup logical storage object as a read and write object for storing a cluster consistent backup of the database; and

mounting clones of the snapshots of all the logical storage objects from the first phase of the backup operation as read only objects.

2. The method of claim 1 , further comprising:

retrieving each node's data directory and renaming data files maintained by each node for storing a portion of the database.

3. The method of claim 2 , further comprising:

executing compaction of the data files for the plurality of nodes by eliminating duplicate data for portions of the database.

4. The method of claim 3 , further comprising:

generating a snapshot of the backup logical storage object after compaction; and

storing token information for the plurality of nodes.

5. The method of claim 4 , further comprising:

deleting clones of the snapshot of the logical storage objects and snapshots of the logical storage objects from the first phase.

6. The method of claim 4 , further comprising:

using the snapshot of the backup logical storage object to restore the database at a target cluster with a topology different from a topology of the distributed, master-less database cluster.

7. The method of claim 6 , wherein the target cluster topology is same as a topology of the distributed, master-less, database cluster.

8. A non-transitory, machine-readable storage medium having stored thereon instructions for performing a method for a distributed, master-less database cluster having a plurality of nodes each executing a database application and storing at least a portion of a database for the cluster, comprising machine executable code which when executed by at least one machine, causes the machine to:

capture a topology of the distributed, master-less database cluster prior to a first phase of a backup operation;

associate token information of each node with a logical storage object of a shared storage system that stores data of the database for each node on behalf of the database application; wherein the token information is based on a hash value of a portion of the data of the database for each node;

generate a snapshot of all logical storage objects associated with each of the plurality of nodes during the first phase;

store token information of each node with backup metadata for the snapshot of all the logical storage objects, where the token information determines what data is stored by each node for the database during the first phase;

initiate a second phase of the backup operation at a recovery node using the backup metadata from the first phase of the backup operation;

clone snapshots of all the logical storage objects from the first phase of the backup operation;

flush information from a commit log for each node to data containers of the database;

create a new backup logical storage object as a read and write object for storing a cluster consistent backup of the database; and

mount clones of the snapshots of all the logical storage objects from the first phase of the backup operation as read only objects.

9. The non-transitory machine-readable storage medium of claim 8 , the machine executable code further causing the machine to:

retrieve each node's data directory and renaming data files maintained by each node for storing a portion of the database.

10. The non-transitory machine-readable storage medium of claim 9 , the machine executable code further causing the machine to:

execute compaction of the data files for the plurality of nodes by eliminating duplicate data for portions of the database.

11. The non-transitory machine-readable of claim 10 , the machine executable code further causing the machine to:

generate a snapshot of the backup logical storage object after compaction; and

store token information for the plurality of nodes.

12. The non-transitory machine-readable of claim 11 , the machine executable code further causing the machine to:

delete clones of the snapshot of the logical storage objects and snapshots of the logical storage objects from the first phase.

13. The non-transitory machine-readable of claim 11 , the machine executable code further causing the machine to:

use the snapshot of the backup logical storage object to restore the database at a target cluster with a topology different from a topology of the distributed, master-less database cluster.

14. The non-transitory machine-readable of claim 13 , wherein the target cluster topology is same as a topology of the distributed, master-less, database cluster.

15. A system, comprising:

a memory containing machine readable medium comprising machine executable code having stored thereon instructions; and a processor coupled to the memory, the processor configured to execute the machine executable code to:

capture a topology of a distributed, master-less database cluster having a plurality of nodes each executing a database application and storing at least a portion of a database for the cluster, prior to a first phase of a backup operation;

associate token information of each node with a logical storage object of a shared storage system that stores data of the database for each node on behalf of the database application; wherein the token information is based on a hash value of a portion of the data of the database for each node;

generate a snapshot of all logical storage objects associated with each of the plurality of nodes during the first phase;

store token information of each node with backup metadata for the snapshot of all the logical storage objects, where the token information determines what data is stored by each node for the database during the first phase;

initiate a second phase of the backup operation at a recovery node using the backup metadata from the first phase of the backup operation;

clone snapshots of all the logical storage objects from the first phase of the backup operation;

flush information from a commit log for each node to data containers of the database;

create a new backup logical storage object as a read and write object for storing a cluster consistent backup of the database; and

mount clones of the snapshots of all the logical storage objects from the first phase of the backup operation as read only objects.

16. The system of claim 15 , the machine executable code further causing the machine to:

retrieve each node's data directory and renaming data files maintained by each node for storing a portion of the database.

17. The system of claim 16 , the machine executable code further causing the machine to:

execute compaction of the data files for the plurality of nodes by eliminating duplicate data for portions of the database.

18. The system of claim 17 , the machine executable code further causing the machine to:

generate a snapshot of the backup logical storage object after compaction; and

store token information for the plurality of nodes.

19. The system of claim 17 , the machine executable code further causing the machine to:

delete clones of the snapshot of the logical storage objects and snapshots of the logical storage objects from the first phase.

20. The system of claim 17 , the machine executable code further causing the machine to:

use the snapshot of the backup logical storage object to restore the database at a target cluster with a topology different from a topology of the distributed, master-less database cluster.