Accelerated data storage synchronization for node fault protection in distributed storage system

Patent metadata

Field	Value
Publication number	US-10244069-B1
Application number	US-201514998102-A
Country	US
Kind code	B1
Filing date	Dec 24, 2015
Priority date	Dec 24, 2015
Publication date	Mar 26, 2019
Grant date	Mar 26, 2019

Abstract

Systems, methods, and articles of manufacture comprising processor-readable storage media are provided for implementing an accelerated data storage synchronization to protect data in a storage system. For example, a first server node accesses a data block from a protected storage system in response to a request from an application executing on the first server node. A modified data block is written to a write-back cache of the first server node. A synchronization operation is performed in response to a request by the application, which includes sending a copy of the modified data block in the write-back cache to a second server node to temporarily store the copy of the modified data block in a storage media of the second server node, and informing the application that the synchronization operation is deemed complete after the copy of the modified data block is stored in the storage media of the second server node.

First claim

What is claimed is:

1. A method for managing memory, comprising:
accessing, by a first server node, a data block from a protected data block storage system in response to a request for the data block from an application executing on the first server node;
writing, by the first server node, a modified data block to a write-back cache that is implemented in system memory residing on the first server node, wherein the modified data block comprises a modified version of the accessed data block which is generated by the application modifying the accessed data block;
protecting the modified data block for write-back to the protected data block storage system by the first server node performing a storage synchronization operation in response to a storage synchronization request by the application, wherein the storage synchronization operation comprises:
the first server node sending a copy of the modified data block in the write-back cache to a second server node to temporarily store the copy of the modified data block in storage media locally residing on the second server node, wherein said storage media is separate from the protected data block storage system;
the first server node informing the application that the storage synchronization operation is deemed complete at least with respect to the modified data block in the write-back cache, in response to the copy of the modified data block being stored in said storage media locally residing on the second server node; and
the first server node marking the modified data block in the write-back cache as being protected, in response to the synchronization operation being deemed complete; and

initiating, by the first server node, an asynchronous write-back operation to store a copy of the modified data block in the write-back cache, which is marked as being protected, in the protected data block storage system.

2. The method of claim 1 , wherein the copy of the modified data block is sent from the first server node to the second server node over a dedicated communication link between the first and second server nodes, which is separate from a communications network used by the first and second server nodes for peer-to-peer communication.

3. The method of claim 1 , further comprising:
marking, by the first server node, the modified data block in the write back cache as unmodified, after the copy of the modified data block is stored in the protected data block storage system; and
causing, by the first server node, the copy of the modified data block in said storage media locally residing on the second server node to be invalidated.

4. The method of claim 1 , wherein the asynchronous write-back operation comprises:
determining at least two locations in the protected data storage system which correspond to the modified data block; and
causing the copy of the modified data block in the write-back cache to be stored in the at least two locations in the protected data storage system.

5. The method of claim l, wherein performing the storage synchronization operation by the first server node further comprises:
utilizing, by the first server node, a remote direct memory access protocol for remotely controlling the storage media locally residing on the second server node;
wherein remotely controlling the storage media comprises:
writing the copy of the modified data block in the storage media locally residing on the second server node; and
invalidating copies of one or more other modified data blocks previously stored in the storage media locally residing on the second server node.

6. The method of claim 1 , wherein performing the storage synchronization operation by the first server node further comprises the first server node sending a message together with the copy of the modified data block to the second server node, wherein the message indicates which copies of one or more other modified data blocks previously stored in the storage media locally residing on the second server node can be invalidated and overwritten.

7. The method of claim 1 , wherein performing the storage synchronization operation by the first server node further comprises:
determining, by the first server node, an available storage capacity of the storage media locally residing on the second server node; and
causing one or more modified data blocks in the write-back cache to be sent to the second server node synchronously with storing all modified data blocks in the write-back cache to corresponding locations in the protected data block storage system;
wherein the one or more modified data blocks sent to the second server node do not exceed the determined available storage capacity of the storage media locally residing on the second server node.

8. The method of claim 1 , wherein the first and second server nodes implement a peer-to-peer server cluster system, and wherein the protected data block storage system comprises at least one of a cluster of DAS (direct attached storage) devices, a SAN (storage area network) comprising a cluster of storage media devices, and a virtual SAN comprising a cluster of storage media devices.

9. The method of claim 1 , wherein said storage media locally residing on the second server node comprises system memory of the second server node.

10. An article of manufacture comprising a processor-readable storage medium having stored therein program code of one or more software programs, wherein the program code is executable by one or more processors to execute a process comprising:
accessing, by a first server node, a data block from a protected data block storage system in response to a request for the data block from an application executing on the first server node;
writing, by the first server node, a modified data block to a write-back cache that is implemented in system memory residing on the first server node, wherein the modified data block comprises a modified version of the accessed data block which is generated by the application modifying the accessed data block;
protecting the modified data block for write-back to the protected data block storage system by the first server node performing a storage synchronization operation in response to a storage synchronization request by the application, wherein the storage synchronization operation comprises:
the first server node sending a copy of the modified data block in the write-back cache to a second server node to temporarily store the copy of the modified data block in storage media locally residing on the second server node, wherein said storage media is separate from the protected data block storage system;
the first server node informing the application that the storage synchronization operation is deemed complete at least with respect to the modified data block in the write-back cache, in response to the copy of the modified data block being stored in said storage media locally residing on the second server node; and
the first server node marking the modified data block in the write-back cache as being protected, in response to the synchronization operation being deemed complete; and

initiating, by the first server node, an asynchronous write-back operation to store a copy of the modified data block in the write-back cache, which is marked as being protected, in the protected data block storage system.

11. The article of manufacture of claim 10 , wherein the copy of the modified data block is sent from the first server node to the second server node over a dedicated communication link between the first and second server nodes, which is separate from a communications network used by the first and second server nodes for peer-to-peer communication.

12. The art

Assignees

• Emc Corp
• Emc Ip Holding Co Llc

Inventors

• Michaud Adrian
• Shain Randall
• Harwood John S
• Taylor Kenneth J
• Au Stephen Wing-Kin

Classifications

• H04L67/104

  Peer-to-peer [P2P] networks · CPC title

• G06F2212/621

  Coherency control relating to peripheral accessing, e.g. from DMA or I/O device · CPC title

• H04L67/1095

  Replication or mirroring of data, e.g. scheduling or transport for data synchronisation between network nodes · CPC title

• G06F2212/154

  Networked environment · CPC title

• G06F12/0815

  Cache consistency protocols · CPC title