Fast migration of metadata

1. A system, comprising:

a first storage node configured to:

determine one or more underlying database files that correspond to one or more selected buckets of key-value pairs;

directly copy to a second storage node the one or more underlying database files that correspond to one or more selected buckets of key-value pairs, wherein updates to the one or more underlying database files are suspended while the one or more underlying database files are being sent to the second storage node; and

the second storage node configured to:

receive the one or more underlying database files; and

update a local key-value store based on the one or more received underlying database files.

2. The system of claim 1 , wherein the first storage node is configured to:

log the updates to the one or more underlying database files after the updates to the one or more database files are suspended while the one or more underlying database files are being sent to the second storage node; and

send the logged updates to the second storage node.

3. The system of claim 2 , wherein the second storage node is configured to apply the logged updates to the one or more received underlying database files.

4. The system of claim 2 , wherein the first storage node is configured to inspect a log after the one or more underlying database files are copied to the second storage node.

5. The system of claim 4 , wherein the first storage node is configured to determine whether a number of entries included in the log is less than a threshold number of updates.

6. The system of claim 5 , wherein in the event the number of entries included in the log is less than the threshold number of updates, the first storage node is configured to:

suspend updates to the one or more selected buckets of key-value pairs; and

send the log to the second storage node.

7. The system of claim 5 , wherein in the event the number of entries included in the log is not less than the threshold number of updates, the first storage node is configured to determine whether a retry threshold has been exceeded.

8. The system of claim 7 , wherein in the event the retry threshold has not been exceeded, the first storage node is configured to:

generate a new log; and

send the log to the second storage node.

9. The system of claim 7 , wherein in the event the retry threshold has been exceeded, the first storage node is configured to:

suspend updates to the one or more selected buckets of key-value pairs; and

send the log to the second storage node.

10. The system of claim 1 , wherein the first storage node is configured to suspend the updates to the one or more database files.

11. The system of claim 1 , wherein the one or more selected buckets of key-value pairs correspond to one or more buckets of key-value pairs that are accessed at a frequency greater than a threshold.

12. The system of claim 1 , wherein the second storage node is a new storage node associated within the system.

13. The system of claim 1 , wherein the local key-value store is a log structured merge database.

14. The system of claim 1 , wherein the first storage node is configured to update a master table that indicates which storage node of a plurality of storage nodes is to handle a file operation for a particular key-value pair.

15. The system of claim 14 , wherein entries included in the master table that correspond to the key-value pairs included in the one or more selected buckets are updated to reference the second storage node in place of the first storage node.

16. A method, comprising:

determining, by a first storage node, one or more underlying database files that correspond to one or more selected buckets of key-value pairs; and

directly copying, from the first storage node to a second storage node, the one or more underlying database files that correspond to one or more selected buckets of key-value pairs,

wherein updates to the one or more underlying database files are suspended while the one or more underlying database files are being sent to the second storage node,

wherein the second storage node receives the one or more underlying database files and updates a local key-value store based on the one or more received underlying database files.

17. The method of claim 16 , wherein the one or more selected buckets of key-value pairs correspond to one or more buckets of key-value pairs that are accessed at a frequency greater than a threshold.

18. The method of claim 16 , further comprising:

logging the updates to the one or more underlying database files after the updates to the one or more database files are suspended while the one or more underlying database files are being sent to the second storage node; and

sending the logged updates to the second storage node.

19. The method of claim 18 , wherein the second storage node applies the logged updates to the one or more received underlying database files.

20. A computer program product embodied in a non-transitory computer readable medium and comprising computer instructions for:

determining, by a first storage node, one or more underlying database files that correspond to one or more selected buckets of key-value pairs; and

directly copying, from the first storage node to a second storage node, the one or more underlying database files that correspond to one or more selected buckets of key-value pairs,

wherein updates to the one or more underlying database files are suspended while the one or more underlying database files are being sent to the second storage node,

wherein the second storage node receives the one or more underlying database files and updates a local key-value store based on the one or more received underlying database files.