Multi-node replication systems, devices and methods

• US 9,158,779 B2
• Filed: 09/14/2012
• Issued: 10/13/2015
• Est. Priority Date: 05/17/2007
• Status: Expired due to Fees

• Current Assignee: Micro Focus Software, Inc. (Open Text Corporation)
• Original Assignee: Novell Incorporated (Open Text Corporation)
• Inventors: Sudhakar, Gosukonda Naga Venkata Satya
• Primary Examiner(s): MORRISON, JAY A
• Application Number:
  US13/619,674
• Publication Number:
  US 20130013562A1
• Time in Patent Office: 1,124 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06F 11/2097 maintaining the standby controller/processing unit updated initialisation or re-synchronisation thereof G06F11/1658 and subgroups
• CPC Class Codes: G06F 16/10 File systems; File servers

Abstract

Replication techniques are presented. According to an embodiment of a method, a node of a replicated storage network is assigned to be an owner of a data block to issue write memory block commands. The network includes at least two nodes including the node assigned to be the owner. If a read memory block command is received to read the data block, a read_lock is issued for the data block, the data block is read, and the read_lock for the data block is released. If a write memory block command is received to write new data to the data block, a write_lock is issued for the data block, the data block is written and a version associated with the data block is incremented, and the write_lock for the data block is released.

Claims

1. A method, comprising:

assigning a replicated storage node of a replicated storage network to be an owner of a replicated data block in the replicated storage network, the owner of the replicated data block to issue write memory block commands to write data to the replicated data block, the network including at least two replicated storage nodes including the replicated storage node assigned to be the owner of the replicated data block, each of the replicated storage nodes to store the replicated data block, the replicated storage network including a plurality of replicated data blocks where each of the plurality of replicated data blocks is individually assigned to be owned by one of the at least two replicated storage nodes;

when a read memory block command is received to read the replicated data block, issuing a read_lock for the replicated data block, reading the replicated data block, and releasing the read_lock for the replicated data block; and

when a write memory block command is received to write new data to the replicated data block, issuing a write_lock for the replicated data block, writing the replicated data block and incrementing a version associated with the replicated data block, and releasing the write_lock for the replicated data block, wherein;

each of the plurality of replicated data blocks being identified by a unique block number;

each replicated storage node of the network is identified by a unique node number; and

assigning a replicated storage node to be the owner of the replicated data block includes dividing the block number of the replicated data block by the number of replicated storage nodes in the network and identifying a resulting remainder value; and

assigning a replicated storage node with a node number that corresponds to the remainder value to be the owner of the replicated data block.

2. The method of claim 1, wherein assigning a replicated storage node to be the owner of the replicated data block includes mapping the plurality of replicated data blocks across all replicated storage nodes in the network.

3. The method of claim 1, wherein:

assigning a replicated storage node to be the owner of the replicated data block includes using a mapping function that maps a given block number to a replicated storage node.

4. The method of claim 1, wherein issuing a read_lock includes issuing a read_lock specific to the replicated data block, and issuing a write_lock includes issuing a write_lock specific to the replicated data block.

5. The method of claim 4, wherein issuing the write_lock includes issuing a distributed write_lock for the network.

6. The method of claim 1, wherein assigning the replicated storage node includes comparing a number of times a replicated storage node acts as an owner node to a number of times other replicated storage nodes act as an owner node to generate a comparison result;

and adjusting assigned ownerships to distribute ownership activity among network replicated storage nodes using the comparison result.

7. The method of claim 1, wherein the method is applied to share resources on the replicating storage network, wherein the resources include disk blocks.

8. The method of claim 1, wherein the method is applied to share resources on the replicating storage network, wherein the resources include files.

9. The method of claim 1, wherein the method is applied to share resources on the replicating storage network, wherein the resources include database records.

10. A multi-node replicated storage network, comprising:

at least two network replicated storage nodes, each network replicated storage node including;

a memory storage device, including replicated storage having a plurality of replicated data blocks stored therein and a version associated with each of the replicated data blocks;

a network link interface; and

a replication manager to communicate with other network replicated storage nodes using the network link interface, and the replication manager to communicate with the memory storage device;

wherein the replication managers for the at least two network replicated storage nodes are adapted to;

determine node ownership for each of the plurality of replicated data blocks each replicated data block being individually owned by one of the replicated storage nodes, the replicated storage node that owns a given one of the replicated data blocks to issue write memory block commands for the given one of the replicated data blocks; and

write a first replicated data block using a first process at an owner replicated storage node for the first replicated data block and using a second process at non-owner replicated storage nodes, wherein the first process at the owner replicated storage node includes;

issuing a write_lock request for the first replicated data block to all replicated storage nodes in the network;

copying old data for the first replicated data block to a memory location, and issuing a write command to non-owner replicated storage nodes in the network to write new data for the first replicated data block;

waiting for the acknowledgment from each non-owner replicated storage node indicating that the non-owner replicated storage node has written the new data for the first replicated data block;

writing the new data for the first replicated data block to the owner replicated storage node, and

incrementing a version associated with the first replicated data block;

issuing a write_unlock request for the first replicated data block to all replicated storage nodes on the network; and

performing a write_unlock for the first replicated data block on the owner replicated storage node.

11. A multi-node replicated storage network, comprising:

at least two network replicated storage nodes, each network replicated storage node including;

a memory storage device, including replicated storage having a plurality of replicated data blocks stored therein and a version associated with each of the replicated data blocks;

a network link interface; and

a replication manager to communicate with other network replicated storage nodes using the network link interface, and the replication manager to communicate with the memory storage device;

wherein the replication managers for the at least two network replicated storage nodes are adapted to;

determine node ownership for each of the plurality of replicated data blocks each replicated data block being individually owned by one of the replicated storage nodes, the replicated storage node that owns a given one of the replicated data blocks to issue write memory block commands for the given one of the replicated data blocks; and

write a first replicated data block using a first process at an owner replicated storage node for the first replicated data block and using a second process at non-owner replicated storage nodes, wherein the second process at non-owner replicated storage nodes includes;

receiving the write_lock request for the first replicated data block from the owner replicated storage node for the first replicated data block;

performing a write_lock for the first replicated data block, and copying old data for the first replicated data block into memory;

writing the new data for the first replicated data block, and incrementing a version number associated with the first replicated data block;

sending the acknowledgment to the owner replicated storage node; and

performing a write_unlock for the first replicated data block.

12. A method, comprising:

assigning a replicated storage node of a replicated storage network to be an owner of a replicated data block in the replicated storage network, the owner of the replicated data block to issue write memory block commands to write data to the replicated data block, the network including at least two replicated storage nodes including the replicated storage node assigned to be the owner of the replicated data block, each of the replicated storage nodes to store the replicated data block, the replicated storage network including a plurality of replicated data blocks where each of the plurality of replicated data blocks is individually assigned to be owned by one of the at least two replicated storage nodes;

when a read memory block command is received to read the replicated data block, issuing a read_lock for the replicated data block, reading the replicated data block, and releasing the read_lock for the replicated data block; and

when a write memory block command is received to write new data to the replicated data block, issuing a write_lock for the replicated data block, writing the replicated data block and incrementing a version associated with the replicated data block, and releasing the write_lock for the replicated data block, wherein;

assigning a replicated storage node to be the owner of the replicated data block includes mapping the plurality of replicated data blocks across all replicated storage nodes in the network, and

mapping the plurality of replicated data blocks includes dividing a block number by a number of the replicated storage nodes in the network, and using a remainder value from dividing to assign the replicated storage node.

13. The method of claim 12, further comprising counting times replicated storage nodes act as an owner of replicated data blocks in the network, and exchanging ownership from a high burdened replicated storage node to a low burdened replicated storage node.

14. A multi-node replicated storage network, comprising:

at least two network replicated storage nodes, each network replicated storage node including;

a memory storage device, including replicated storage having a plurality of replicated data blocks stored therein and a version associated with each of the replicated data blocks;

a network link interface; and

a replication manager to communicate with other network replicated storage nodes using the network link interface, and the replication manager to communicate with the memory storage device;

wherein the replication managers for the at least two network replicated storage nodes are adapted to;

determine node ownership for replicated data blocks to issue write memory block commands, each replicated data block being owned by one replicated storage node; and

write a first replicated data block using a first process at an owner replicated storage node for the first replicated data block and using a second process at non-owner replicated storage nodes,

wherein the first process at the owner replicated storage node includes;

issuing a write_lock request for the first replicated data block to all replicated storage nodes in the network;

copying old data for the first replicated data block to a memory location, and issuing a write command to non-owner replicated storage nodes in the network to write new data for the first replicated data block;

waiting for the acknowledgment from each non-owner replicated storage node indicating that the non-owner replicated storage node has written the new data for the first replicated data block;

writing the new data for the first replicated data block to the owner replicated storage node;

incrementing a version associated with the first replicated data block;

issuing a write_unlock request for the first replicated data block to all replicated storage nodes on the network; and

performing a write_unlock for the first replicated data block on the owner replicated storage node.

15. A multi-node replicated storage network, comprising:

at least two network replicated storage nodes, each network replicated storage node including;

a memory storage device, including replicated storage having a plurality of replicated data blocks stored therein and a version associated with each of the replicated data blocks;

a network link interface; and

a replication manager to communicate with other network replicated storage nodes using the network link interface, and the replication manager to communicate with the memory storage device;

wherein the replication managers for the at least two network replicated storage nodes are adapted to;

determine node ownership for replicated data blocks to issue write memory block commands, each replicated data block being owned by one replicated storage node; and

write a first replicated data block using a first process at an owner replicated storage node for the first replicated data block and using a second process at non-owner replicated storage nodes,

wherein the second process at non-owner replicated storage nodes includes;

receiving the write_lock request for the first replicated data block from the owner replicated storage node for the first replicated data block;

performing a write_lock for the first replicated data block, and copying old data for the first replicated data block into memory;

writing the new data for the first replicated data block, and incrementing a version number associated with the first replicated data block;

sending the acknowledgment to the owner replicated storage node; and

performing a write_unlock for the first replicated data block.

16. A method, comprising:

assigning a replicated storage node of a replicated storage network to be an owner of a replicated data block to issue write memory block commands, the network including at least two replicated storage nodes including the replicated storage node assigned to be the owner, each of the replicated storage nodes to store the replicated data block;

when a read memory block command is received to read the replicated data block, issuing a read_lock for the replicated data block, reading the replicated data block, and releasing the read_lock for the replicated data block; and

when a write memory block command is received to write new data to the replicated data block, issuing a write_lock for the replicated data block, writing the replicated data block and incrementing a version associated with the replicated data block, and releasing the write_lock for the replicated data block,

wherein the method further comprises;

determining if a network error occurred prior to writing the replicated data block to the owner replicated storage node;

writing the data block to the owner replicated storage node if a network error did not occur; and

if a network error occurred, performing corrective action,

wherein, if a network error occurred prior to writing the data block to the owner replicated storage node, the method further comprises;

determining if the owner replicated storage node is active or passive after performing corrective action; and

aborting the local write of the replicated data block to the owner replicated storage node if the owner replicated storage node is passive after performing corrective action, or writing the replicated data block to the owner replicated storage node if the owner replicated storage node is active after performing the corrective action.

17. The method of claim 16, further comprising making the owner replicated storage node passive after a network error if the owner replicated storage node is part of a cluster of replicated storage nodes that remain in communication with each other that is less than half of the replicated storage nodes of the replicated storage network before the network error.

18. The method of claim 16, wherein determining if the owner replication storage node is active or passive includes an administrator determining if the owner replication storage node is active or passive.