System and method of collision management in a namespace of a storage system
In a namespace distributed across storage devices, collisions of objects stored in the namespace can occur. To manage collisions, first, a collision flag is provided in system metadata of the objects, and second, a predetermined naming convention is applied to the pathnames of collided objects within the namespace. A winner object of a collision is stored with the pathname of the actual object (e.g., “/object”) in the namespace, and the loser object is stored with a pathname derived from the actual object (e.g., “/object.collision”) and the collision flag of the loser object is set to mark the loser object within the namespace. Accordingly, winners and losers of collisions in the namespace can be identified by searching system metadata for collision flags which indicate collision losers, and associated winners can be identified based on the pathname of the losers.
1. A system comprising:
a plurality of storage devices; and
one or more processors coupled to the storage devices, and configured to:
logically partition storage space provided by the storage devices into a namespace which stores a plurality of objects each including data and metadata;
determine, from two objects of the plurality of objects which have collided in the namespace, a winner object and a loser object, and mark the loser object of the collision with a collision flag included in the metadata;
receive a query for collisions in the namespace; and
return, in response to receiving the query, the loser object based on the collision flag,
wherein the one or more processors coupled to the storage devices are further configured to:
set a new pathname for the loser object; and
clear, in response to setting the new pathname for the loser object, the collision flag included in the metadata of the loser object.
2. The system according to claim 1 , wherein the one or more processors coupled to the storage devices are further configured to:
modify a pathname of the winner object with which the loser object collided;
clear, in response to setting the new pathname for the loser object and modifying the pathname of the winner object, the collision flag included in the metadata of the loser object; and
mark, in response to clearing the collision flag included in the metadata of the loser object, the winner object of the collision with a collision flag included in the metadata of the winner object.
3. The system according to claim 1 , wherein the one or more processors coupled to the storage devices are further configured to:
determine the winner object and the loser object based on timestamps of the two objects which have collided, hash values of the two objects which have collided, or storage locations of the two objects which have collided.
4. The system according to claim 1 , wherein the one or more processors coupled to the storage devices are further configured to:
set the new pathname of the loser object to indicate a pathname of the winner object with which the loser object collided, and
identify the winner object based on the pathname of the loser object.
5. The system according to claim 1 , wherein the one or more processors coupled to the storage devices are further configured to:
set the new pathname of the loser object to indicate a pathname of the winner object with which the loser object collided, and
store the loser object and the winner object in a same directory of the namespace.
6. The system according to claim 1 , wherein the one or more processors coupled to the storage devices are further configured to:
set the new pathname of the loser object to indicate a pathname of the winner object with which the loser object collided, and
move the loser object into a predetermined directory of the namespace different from a directory of the winner object.
7. The system according to claim 1 , wherein the one or more processors coupled to the storage devices are further configured to:
return, in response to receiving the query, the loser object based on the collision flag and further return, in response to receiving the query, a pathname, a number of collisions, and a change time of the loser object.
8. The system according to claim 1 , wherein the one or more processors coupled to the storage devices are further configured to:
replicate the metadata of the loser object, including the collision flag, to other storage devices partitioned into the namespace.
9. A system comprising:
a plurality of storage devices; and
one or more processors coupled to the storage devices, and configured to:
logically partition storage space provided by the storage devices into a namespace which stores a plurality of objects each including data and metadata;
determine, from two objects of the plurality of objects which have collided in the namespace, a winner object and a loser object, and mark the loser object of the collision with a collision flag included in the metadata;
receive a query for collisions in the namespace; and
return, in response to receiving the query, the loser object based on the collision flag,
wherein the one or more processors coupled to the storage devices are further configured to:
modify the pathname of the loser object to indicate a pathname of the winner object with which the loser object collided, and
identify the winner object based on the pathname of the loser object.
10. The system according to claim 9 , wherein the one or more processors coupled to the storage devices are further configured to:
clear, in response to modifying the pathname for the loser object, the collision flag included in the metadata of the loser object.
11. The system according to claim 9 , wherein the one or more processors coupled to the storage devices are further configured to:
modify a pathname of the winner object with which the loser object collided;
clear, in response to modifying the pathname for the loser object and modifying the pathname of the winner object, the collision flag included in the metadata of the loser object; and
mark, in response to clearing the collision flag included in the metadata of the loser object, the winner object of the collision with a collision flag included in the metadata of the winner object.
12. The system according to claim 9 , wherein the one or more processors coupled to the storage devices are further configured to:
determine the winner object and the loser object based on timestamps of the two objects which have collided, hash values of the two objects which have collided, or storage locations of the two objects which have collided.
13. The system according to claim 9 , wherein the one or more processors coupled to the storage devices are further configured to:
store the loser object and the winner object in a same directory of the namespace.
14. The system according to claim 9 , wherein the one or more processors coupled to the storage devices are further configured to:
move the loser object into a predetermined directory of the namespace different from a directory of the winner object.
15. The system according to claim 9 , wherein the one or more processors coupled to the storage devices are further configured to:
return, in response to receiving the query, the loser object based on the collision flag and further return, in response to receiving the query, a pathname, a number of collisions, and a change time of the loser object.
16. The system according to claim 9 , wherein the one or more processors coupled to the storage devices are further configured to:
replicate the metadata of the loser object, including the collision flag, to other storage devices partitioned into the namespace.
17. A system comprising:
a plurality of storage devices; and
one or more processors coupled to the storage devices, and configured to:
logically partition storage space provided by the storage devices into a namespace which stores a plurality of objects each including data and metadata;
determine, from two objects of the plurality of objects which have collided in the namespace, a winner object and a loser object, and mark the loser object of the collision with a collision flag included in the metadata;
receive a query for collisions in the namespace; and
return, in response to receiving the query, the loser object based on the collision flag and further return, in response to receiving the query, a pathname, a number of collisions, and a change time of the loser object.
18. The system according to claim 17 , wherein the one or more processors coupled to the storage devices are further configured to:
determine the winner object and the loser object based on timestamps of the two objects which have collided, hash values of the two objects which have collided, or storage locations of the two objects which have collided.
19. The system according to claim 17 , wherein the one or more processors coupled to the storage devices are further configured to:
replicate the metadata of the loser object, including the collision flag, to other storage devices partitioned into the namespace.