Method and system for global namespace with consistent hashing
A method for writing data to a storage pool. The method includes receiving a virtual identifier (ID) and an offset for an object, extracting a node identifier (ID) that identifies a first storage server in the storage pool from the virtual ID, obtaining an object layout (OL) for the object from the first storage server, hashing an object ID and an offset ID to obtain a hashed value, where the virtual ID comprises the object ID and where the offset ID is generated from the offset, identifying a second storage server in the storage pool using a global namespace layout (GL), the OL, and the hashed value, and issuing a write request to write data to the object in the second storage server, where the write request comprises the object ID and the offset.
1. A method for writing data to a storage pool, comprising:
creating a global namespace layout (GL), wherein the GL comprises a first server entry for a first storage server in the storage pool and a second server entry from a second storage server in the storage pool, wherein the first server entry comprises a node ID identifying the first storage server, a node identifier (ID) associated with the first storage server, and a node point used to determine which write requests are serviced by the first storage server;
providing the GL to a client;
after providing the GL to the client:
receiving, from the client, a request for an object layout (OL) by the first storage server;
providing the OL to the client;
and
after sending the OL to the client, receiving by the second storage server, a write request to write data to an object in the second storage server, wherein the write request comprises an object ID and an offset; and
writing in response to the write request, the data to the object on the second storage server.
2. The method of claim 1 , wherein the storage pool comprises a plurality of storage servers, wherein the GL specifies which of the plurality of storage servers were writeable for each GL sequence number and how writes were distributed across the plurality of storage servers writeable for each GL sequence number.
3. The method of claim 1 ,
wherein the OL specifies an OL sequence number, wherein the OL sequence number is set when the object is created, wherein the OL sequence number corresponds to a GL sequence number of the GL when the object was created.
4. The method of claim 1 , wherein the request for the OL comprises a value generated using a virtual identifier (ID) associated with the object.
5. The method of claim 4 , wherein the OL comprises an OL sequence number, the virtual ID, and a stripe width.
6. The method of claim 1 , wherein the GL further comprises:
a first node sequence number corresponding to a GL sequence number when the first storage server was added to the storage pool.
7. The method of claim 1 , further comprising:
receiving, by the second storage server, a request to store the OL in the second storage server, wherein the request to store the OL specifies a virtual ID and a special offset; and
storing, by the second storage server, the OL in the second storage server.
8. The method of claim 1 , further comprising:
providing the GL to a second client:
after providing the GL to the second client:
receiving, by the first storage server, a second GL from the second client, wherein the second GL is associated with a second GL sequence number that is greater than the first GL sequence number associated with the GL; and
storing the second GL in the first storage server.
9. The method of claim 1 , further comprising:
receiving, by the second storage server, a request to delete the object;
in response to receiving the request to delete the object:
setting a delete flag for an entry corresponding to the object in the second storage server;
invalidating all other entries corresponding to the object in all other storage servers in the storage pool; and
after all the other entries corresponding to the object in all other storage servers in the storage pool have been invalidated, invalidating the entry corresponding to the object in the second storage server.
10. A non-transitory computer readable medium comprising instructions which when executed by a processor perform a method, the method comprising:
creating a global namespace layout (GL), wherein the GL comprises a first server entry for a first storage server in the storage pool and a second server entry from a second storage server in the storage pool, wherein the first server entry comprises a node ID identifying the first storage server, a node identifier (ID) associated with the first storage server, and a node point used to determine which write requests are serviced by the first storage server;
providing the GL to a client;
after providing the GL to the client:
receiving, from the client, a request for an object layout (OL) by the first storage server;
providing the OL to the client; and
after sending the OL to the client, receiving by the second storage server, a write request to write data to an object in the second storage server, wherein the write request comprises an object ID and an offset; and
writing in response to the write request, the data to the object on the second storage server.
11. The non-transitory computer readable medium of claim 10 , wherein the storage pool comprises a plurality of storage servers, wherein the GL specifies which of the plurality of storage servers were writeable for each GL sequence number and how writes were distributed across the plurality of storage servers writeable for each GL sequence number.
12. The non-transitory computer readable medium of claim 10 , wherein the OL specifies an OL sequence number, wherein the OL sequence number is set when the object is created, wherein the OL sequence number corresponds to a GL sequence number of the GL when the object was created.
13. The non-transitory computer readable medium of claim 10 , wherein the request for the OL comprises a value generated using a virtual identifier (ID) associated with the object.
14. The non-transitory computer readable medium of claim 13 , wherein the OL comprises an OL sequence number, the virtual ID, and a stripe width.
15. The non-transitory computer readable medium of claim 10 , wherein the GL further comprises: a first node sequence number corresponding to a GL sequence number when the first storage server was added to the storage pool.
16. The non-transitory computer readable medium of claim 10 , wherein the method further comprises:
receiving, by the second storage server, a request to store the OL in the second storage server, wherein the request to store the OL specifies a virtual ID and a special offset; and
storing, by the second storage server, the OL in the second storage server.
17. The non-transitory computer readable medium of claim 10 , wherein the method further comprises:
providing the GL to a second client:
after providing the GL to the second client:
receiving, by the first storage server, a second GL from the second client, wherein the second GL is associated with a second GL sequence number that is greater than the first GL sequence number associated with the GL; and
storing the second GL in the first storage server.
18. The non-transitory computer readable medium of claim 10 , wherein the method further comprises:
receiving, by the second storage server, a request to delete the object;
in response to receiving the request to delete the object:
setting a delete flag for an entry corresponding to the object in the second storage server;
invalidating all other entries corresponding to the object in all other storage servers in the storage pool; and
after all the other entries corresponding to the object in all other storage servers in the storage pool have been invalidated, invalidating the entry corresponding to the object in the second storage server.
19. A system, comprising:
a first storage server comprising a first persistent storage; and
a second storage server comprising a second persistent storage;
wherein the first storage server is programmed to:
create a global namespace layout (GL), wherein the GL comprises a first server entry for a first storage server in the storage pool and a second server entry from a second storage server in the storage pool, wherein the first server entry comprises a node ID identifying the first storage server, a node identifier (ID) associated with the first storage server, and a node point used to determine which write requests are serviced by the first storage server;
after providing the GL to the client:
provide the GL to a client;
receive, from the client, a request for an object layout (OL) by the first storage server;
provide the OL to the client;
wherein the second storage server is programmed to:
after sending the OL to the client, receive by the second storage server, a write request to write data to an object in the second storage server, wherein the write request comprises an object ID and an offset; and
write in response to the write request, the data to the object in the second persistent storage.
20. The system of claim 19 , wherein the first storage server is further programmed to:
provide the GL to a second client:
after providing the GL to the second client:
receive a second GL from the second client, wherein the second GL is associated with a second GL sequence number that is greater than the first GL sequence number associated with the GL; and
storing the second GL.