Resource allocation for synthetic backups

1. A storage system comprising:

a processor; and

a machine-readable storage storing instructions, the instructions executable by the processor to:

determine to generate a synthetic full backup based on data stream representations of a plurality of data streams;

in response to a determination to generate the synthetic full backup, create a logical group including the data stream representations;

specify a cache resource allocation for the logical group; and

generate the synthetic full backup from data stream representations using an amount of a cache resource limited by the cache resource allocation for the logical group.

2. The storage system of claim 1 , including instructions executable by the processor to:

specify a resource policy for the logical group; and

during generation of the synthetic full backup, manage use of the resource allocation within the logical group according to the resource policy.

3. The storage system of claim 2 , wherein the resource policy is one selected from a fixed allocation policy, a bounded scaling policy, a non-linear scaling policy, and a client directed policy.

4. The storage system of claim 1 , wherein the plurality of data streams corresponds to a full backup and a plurality of incremental backups.

5. The storage system of claim 1 , further comprising

a cache; and

a cache controller to:

in response to a detection of a failure in the cache, identify a data stream representation associated with the largest amount of data read from the cache in a first time period;

determine a most recently used cache entry associated with the identified data stream representation; and

promote the determined most recently used cache entry to a position of the cache that is last in eviction order.

6. The storage system of claim 5 , the cache controller to:

track a total number of misses that occur in the cache during the first time period; and

detect the failure in the cache in response to a determination that the total number of misses exceeds a threshold level.

7. The storage system of claim 1 , including instructions executable by the processor to:

in response to a determination that a particular data stream representation is no longer needed to generate the synthetic full backup, remove the particular data stream representation from the logical group.

8. The storage system of claim 1 , including instructions executable by the processor to:

in response to a determination that all of the data stream representations are no longer needed to generate the synthetic full backup, delete the logical group.

9. A non-transitory machine-readable medium storing instructions that upon execution cause a processor to:

determine to generate a synthetic full backup based on data stream representations of a plurality of data streams;

in response to a determination to generate the synthetic full backup, create a logical group including the data stream representations;

assign a resource policy and a resource allocation to the logical group; and

generate the synthetic full backup from the data stream representations based on the resource policy and the resource allocation assigned to the logical group.

10. The non-transitory machine-readable medium of claim 9 , including instructions that upon execution cause the processor to:

during generation of the synthetic full backup, manage use of the resource allocation within the logical group according to the resource policy.

11. The non-transitory machine-readable medium of claim 9 , wherein the resource policy is one selected from a fixed allocation policy, a bounded scaling policy, a non-linear scaling policy, and a client directed policy.

12. The non-transitory machine-readable medium of claim 9 , including instructions that upon execution cause the processor to:

in response to a determination that a particular data stream representation is no longer needed to generate the synthetic full backup, remove the particular data stream representation from the logical group.

13. The non-transitory machine-readable medium of claim 9 , including instructions that upon execution cause the processor to:

in response to a determination that all of the data stream representations are no longer needed to generate the synthetic full backup, delete the logical group.

14. A method comprising:

determining, by a controller of a storage device, whether to generate a synthetic full backup based on data stream representations of a plurality of data streams;

in response to a determination to generate the synthetic full backup based on the data stream representations of the plurality of data streams, creating, by the controller, a logical group including the data stream representations of the plurality of data streams;

assigning, by the controller, a resource allocation to the logical group including the data stream representations of the plurality of data streams; and

generating, by the controller, the synthetic full backup from the data stream representations based on the resource allocation assigned to the logical group.

15. The method of claim 14 , comprising:

specifying a resource policy for the logical group; and

during generation of the synthetic full backup, managing use of the resource allocation within the logical group according to the resource policy.

16. The method of claim 15 , wherein the resource policy is one selected from a fixed allocation policy, a bounded scaling policy, a non-linear scaling policy, and a client directed policy.

17. The method of claim 14 , wherein the plurality of data streams corresponds to a full backup and a plurality of incremental backups.

18. The method of claim 14 , comprising:

in response to a determination that a particular data stream representation is no longer needed to generate the synthetic full backup, removing the particular data stream representation from the logical group.

19. The method of claim 14 , comprising:

in response to a determination that all of the data stream representations are no longer needed to generate the synthetic full backup, deleting the logical group.

20. The method of claim 14 , comprising:

detecting a failure in a cache in response to a determination that a total number of misses exceeds a threshold level.