SYSTEM AND METHOD FOR NEGOTIATED TAKEOVER OF STORAGE OBJECTS

• US 20150149416A1
• Filed: 11/22/2013
• Published: 05/28/2015
• Est. Priority Date: 11/22/2013
• Status: Active Grant

• Current Assignee: NetApp, Inc.
• Original Assignee: NetApp, Inc.
• Inventors: DHAVALE, Poonam, Bhuwania, Manish, Ramachandran, Arun, Mutiso, Herman M.
• Granted Patent: US 9,424,301 B2
• Time in Patent Office: Days
• US Class Current: 707/687
• CPC Class Codes: G06F 16/2308 Concurrency control transaction processing G06F9/466

Abstract

A system and method of negotiated takeover of storage objects includes one or more processors, a storage controller, and memory coupled to the one or more processors. The memory stores a data structure that includes information about a plurality of storage objects manageable by the storage controller. The storage controller is configured to assume, one by one, current ownership of a first subset of the storage objects and assume, concurrently, current ownership of a second subset of the storage objects. The first subset of storage objects and the second subset of storage objects are currently owned by a second storage server coupled to the storage server. In some embodiments, current ownership of the first subset of storage objects is transferred by iteratively detecting a particular storage object from the first subset of the storage objects whose current ownership can be assumed and bringing the particular storage object online.

Claims

1. A storage server comprising:

one or more processors;

a storage controller; and

memory coupled to the one or more processors and storing a data structure including information about a plurality of storage objects manageable by the storage controller;

wherein the storage controller is configured to;

assume, one by one, current ownership of a first subset of the storage objects, the first subset of storage objects being currently owned by a second storage server coupled to the storage server; and

assume, concurrently, current ownership of a second subset of the storage objects, the second subset of the storage objects being currently owned by the second storage server.

2. The storage server of claim 1 wherein the storage controller is further configured to assume current ownership of the first subset of the storage objects by:

sending a takeover message to the second storage server;

detecting a particular storage object from the first subset of the storage objects whose current ownership can be assumed;

bringing the particular storage object online; and

iteratively detecting and bringing online additional storage objects until the second storage server indicates that there are no more storage objects within the first subset to be taken over.

3. The storage server of claim 2 wherein the storage controller is further configured to receive a source done message from the second storage server to detect the particular storage object, the source done message including an identifier associated with the particular storage object.

4. The storage server of claim 2 wherein the storage controller is further configured to receive a relocation done message from the second storage server to indicate that there are no more storage objects within the first subset to be taken over.

5. The storage server of claim 2 wherein the storage controller is further configured to send a storage object online message to the second storage server after bringing the particular storage object online, the storage object online message including an identifier associated with the particular storage object.

6. The storage server of claim 1 wherein the storage controller is further configured to assume current ownership of the second subset of the storage objects by:

sending a shutdown message to the second storage server;

receiving a shutdown done message from the second storage server;

detecting the second subset of the storage objects by scanning the data structure of storage objects;

changing current ownership of the second subset of the storage objects to the storage server; and

bringing each storage object in the second subset of the storage objects online.

7. The storage server of claim 6 wherein the storage controller is further configured to:

start a timer after sending the shutdown message; and

detect the second subset of the storage objects when the timer times out.

8. The storage server of claim 1 wherein the storage controller is further configured to:

transfer, one by one, current ownership of a third subset of the storage objects, the current ownership being transferred to a third storage server coupled to the storage server; and

transfer, concurrently, current ownership of a fourth subset of the storage objects, the current ownership being transferred to the third storage server.

9. The storage server of claim 8 wherein the third storage server and the second storage server are the same.

10. A non-transitory machine-readable medium comprising a first plurality of machine-readable instructions which when executed by one or more processors associated with a source storage controller cause the one or more processors to perform a method comprising:

receiving a takeover message from a target storage controller;

transferring, one by one, current ownership of a first plurality of storage objects, the current ownership being transferred from the source storage controller to the target storage controller;

sending a relocation done message to the target storage controller;

receiving a shutdown message from the target storage controller; and

giving up, concurrently, current ownership of a second plurality of storage objects.

11. The non-transitory machine-readable medium of claim 10 wherein transferring current ownership of the first plurality of storage objects comprises:

selecting a particular storage object from the first plurality of storage objects;

taking the particular storage object offline;

changing current ownership of the particular storage object to the target storage controller;

detecting that the target storage controller has assumed current ownership of the particular storage object; and

iteratively selecting, taking offline, and changing current ownership of any additional storage objects in the first plurality of storage objects.

12. The non-transitory machine-readable medium of claim 11 wherein detecting that the target storage controller has assumed current ownership of the particular storage object comprises receiving a storage object online message from the target storage controller, the storage object online messaging including an identifier associated with the particular storage object.

13. The non-transitory machine-readable medium of claim 11 wherein detecting that the target storage controller has assumed current ownership of the particular storage object comprises detecting that the particular storage object has been brought online.

14. The non-transitory machine-readable medium of claim 10 wherein giving up current ownership of the second plurality of storage objects comprises:

taking each of the second plurality of storage objects offline; and

sending a shutdown done message to the target storage controller.

15. A method of managing storage objects in a first storage server, the method comprising:

entering a serial takeover phase;

sending a first takeover message to a second storage server;

waiting for the second storage server to take a first storage object offline and transfer a current ownership of the first storage object to the first storage server;

assuming current ownership of the first storage object by bringing the first storage object online;

receiving a first relocation done message from the second storage server;

entering a parallel takeover phase;

detecting a second storage object and a third storage object whose current ownership can be assumed, wherein the second storage object and the third storage object are concurrently offline; and

assuming current ownership of the second storage object and the third storage object by bringing the second storage object online and the third storage object online.

16. The method of claim 15 wherein during the serial takeover phase the method further comprises:

waiting for the second storage server to take a fourth storage object offline and transfer a current ownership of the fourth storage object to the first storage server;

assuming current ownership of the fourth storage object by bringing the fourth storage object online;

17. The method of claim 15, further comprising receiving a source done message from the second storage server before assuming current ownership of the first storage object, the source done message including an identifier associated with the first storage object.

18. The method of claim 15, further comprising sending a storage object online message to the second storage server after assuming ownership of the first storage object, the storage object online message including an identifier associated with the first storage object.

19. The method of claim 15, further comprising sending a shutdown message to the second storage server after entering the parallel takeover phase.

20. The method of claim 15, further comprising:

receiving a second takeover message from a third storage server;

taking a fourth storage object offline;

transferring current ownership of the fourth storage object to the third storage server;

waiting for the third storage server to bring the fourth storage object online;

taking a fifth storage object offline;

transferring current ownership of the fifth storage object to the third storage server;

waiting for the third storage server to bring the fifth storage object online;

sending a second relocation done message to the third storage server;

taking a sixth storage object and an seventh storage object offline, wherein the sixth storage object and the seventh storage object are concurrently offline; and

sending a second shutdown done message to the third storage server.