Distributed storage network memory access based on memory state

• US 8,473,677 B2
• Filed: 05/11/2010
• Issued: 06/25/2013
• Est. Priority Date: 09/29/2009
• Status: Expired due to Fees

• Current Assignee: Pure Storage, Inc.
• Original Assignee: Cleversafe Incorporated (International Business Machines Corporation)
• Inventors: Gladwin, S. Christopher, Leggette, Wesley
• Primary Examiner(s): PATEL, KAUSHIKKUMAR M
• Application Number:
  US12/777,904
• Publication Number:
  US 20110078372A1
• Time in Patent Office: 1,141 Days
• Field of Search: 711/114, 711/153, 711/162, 711/164
• US Class Current: 711/114
• CPC Class Codes: G06F 11/1076 Parity data used in redundant arrays of independent storages, e.g. in RAID systems
• CPC Class Codes: G06F 12/0813 with a network or matrix configuration
• CPC Class Codes: G06F 16/182 Distributed file systems
• CPC Class Codes: G06F 2211/1011 Clustered RAID, i.e. clustered or de-clustered RAID where data and parity are spread over more disks than blocks in a parity group
• CPC Class Codes: G06F 3/0617 in relation to availability
• CPC Class Codes: G06F 3/0656 Data buffering arrangements
• CPC Class Codes: G06F 3/0659 Command handling arrangements, e.g. command buffers, queues, command scheduling
• CPC Class Codes: G06F 3/067 Distributed or networked storage systems, e.g. storage area networks [SAN], network attached storage [NAS]
• CPC Class Codes: G06F 3/0689 Disk arrays, e.g. RAID, JBOD
• CPC Class Codes: H04L 67/1008 based on parameters of servers, e.g. available memory or workload monitoring of computer activity G06F11/30
• CPC Class Codes: H04L 67/101 based on network conditions
• CPC Class Codes: H04L 67/1097 for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS]
• CPC Class Codes: H04N 21/231 Content storage operation, e.g. caching movies for short term storage, replicating data over plural servers, prioritizing data for deletion
• CPC Class Codes: H04N 21/2405 Monitoring of the internal components or processes of the server, e.g. server load allocation of resources in multiprogramming arrangements G06F9/50; performance measurement of computer activity G06F11/34
• CPC Class Codes: H04N 21/442 Monitoring of processes or resources, e.g. detecting the failure of a recording device, monitoring the downstream bandwidth, the number of times a movie has been viewed, the storage space available from the internal hard disk arrangements for monitoring broadcast services or broadcast-related services H04H60/29; arrangements for identifying or recognising characteristics with a direct linkage to broadcast information H04H60/35; monitoring of user activities for profile generation for accessing a video database G06F16/739; monitoring in wireless networks H04W24/00

Abstract

A distributed storage unit determines how to handle a read or write request for a data slice based on a state of the memory the data slice is to be read from or written to. When receiving a request to retrieve a data slice, the distributed storage unit, determines a state of the memory in which the data slice is stored. Based on the memory state, one of multiple different methods for obtaining the data slice is selected. The methods include, among others, a direct read from the memory, and reconstructing the data slice using other memories and parity values. In response to a write request, the distributed storage unit can determine whether to use the currently selected memory for writing, or rotate the memory used for writing, based on a state of the memory.

Claims

1. A method for execution by a distributed storage (DS) unit of a distributed storage network, the method comprising:

maintaining memory states of a plurality of memories of the DS unit, wherein a memory of the plurality of memories is in a write only state, a read state, or a write with read priority state; and

receiving a first request regarding a first data slice;

determining a first memory of the plurality of memories associated with the first data slice;

determining the memory state of the first memory from the memory states of the plurality of memories;

when the memory state of the first memory is the write only state, the first request is a read request, and the DS unit determines to process the first request;

reconstructing the first data slice from corresponding parity data stored in a parity memory of the DS unit and one or more corresponding data slices stored in one or more other memories of the plurality of memories that are in the read state of the write with read priority state; and

outputting the reconstructed encoded data slice in response to the first request

when the memory state of the first memory is write with read priority and the first request is the read request;

pausing write requests to the first memory upon receiving the first request to retrieve the first data slice;

retrieving the first data slice from the first memory; and

outputting the retrieved first data slice.

2. The method of claim 1 further comprising:

when the memory state of the first memory is the read state;

retrieving the first data slice from the first memory; and

outputting the retrieved first data slice in response to the first request.

3. The method of claim 1 further comprising:

determining not to process the first request by determining that other distributed storage units of the distributed storage network should be used to reconstruct a data object, wherein the data object was encoded to produce a set of encoded data slices, which includes the first data slice, and

sending a notification denying the first request.

4. The method of claim 1, further comprising:

when the first request is a write request;

identifying the first memory as a primary write memory, wherein the first memory has the write only state and other memories of the plurality of memories are in the read state or the write with read priority state;

writing the first data slice to the first memory;

generating parity information from the first data slice and one or more corresponding data slices stored in one or more other memories of the plurality of memories; and

writing the parity information into the parity memory of the DS unit.

5. A distributed storage unit comprising:

a plurality of memories operable for storing encoded data slices;

one or more parity memories operable for storing parity data of two or more encoded data slices stored in two or more memories of the plurality of memories;

a storage unit control module operable to;

receive a request to retrieve a stored encoded data slice;

determine a memory of the plurality of memories storing the stored encoded data slice

determine a memory state of the memory;

when the memory state is a write only memory state;

retrieve corresponding parity data from the one or more parity memories;

retrieve one or more corresponding encoded data slices of the two or more encoded data slices from the two or more memories;

reconstruct the stored encoded data slice from the corresponding parity data and the one or more corresponding encoded data slices; and

output the reconstructed encoded data slice in response to the request when the memory state is write with read priority;

pause write requests upon receiving the request to retrieve the stored encoded data slice;

retrieve the stored encoded data slice from the memory; and

output the retrieved encoded data slice.

6. The distributed storage unit of claim 5, further comprising:

the storage unit control module to;

maintain memory states of the plurality of memories; and

determine the memory state of the memory from the memory states of the plurality of memories.

7. The distributed storage unit of claim 5, further comprising:

the storage unit control module operable to;

determine that other distributed storage units should be used to reconstruct a data object, wherein the data object was encoded to produce a set of encoded data slices, which includes the stored encoded data slices; and

send a notification denying the request.

8. The distributed storage unit of claim 7, further comprising:

the storage unit control module operable to;

determine that the distributed storage unit does not have enough currently available resources to produce the reconstructed encoded data slice; and

send a notification denying the request.

9. The distributed storage unit of claim 5, further comprising:

the storage unit control module operable to;

when the memory state is read only, retrieve the stored encoded data slice from the memory; and

output the retrieved encoded data slice.

10. The distributed storage unit of claim 5, further comprising:

the storage unit control module operable to, for the memory;

establish the write only state when a write queue for the memory is at or above a high threshold;

establish a read only state when a read queue for the memory is not empty and the write queue is below the high threshold; and

establish a write with read priority when the read queue is empty and the write queue is below the high threshold.

11. A distributed storage unit comprising:

a plurality of memories operable for storing encoded data slices;

one or more parity memories operable for storing parity data of two or more encoded data slices stored in two or more memories of the plurality of memories;

a storage unit control module operable to;

receive a request to write an encoded data slice;

select one of the plurality of memories as a primary memory to store the encoded data slice;

set a state of the selected one of the plurality of memories to a write only memory state of the primary memory and states of other memories of the plurality of memories are to a read only state or a write with read priority state;

store parity information associated with the encoded data slice in one of the one or more parity memories;

store the encoded data slice in the primary memory; and

determine when to rotate the primary memory from the selected one of the plurality of memories to another one of the plurality of memories.

12. The distributed storage unit of claim 11, wherein the storage unit control module is further to:

determine not to rotate the primary memory when a number of queued write requests for the primary memory is greater than an upper threshold; and

determine to rotate the primary memory when the number of queued write requests is less than the upper threshold limit and there are queued read requests for the primary memory.

13. The distributed storage unit of claim 11, further comprising:

the storage unit control module operable to determine to rotate the primary memory to the other memory based on one or more of an expiration of time, a command, a memory utilization indicator, a read request history indicator for the primary memory, and memory performance optimization; and

when the primary memory is rotated to the other memory, set the state of the selected one of the plurality of memories to READ.

14. The distributed storage unit of claim 11, further comprising:

the one or more parity memories includes a plurality of parity memories;

the storage unit control module operable to;

select the one of the one or more parity memories as a primary parity memory; and

determine when to rotate the primary parity memory from the one of the one or more parity memories to another one of the plurality of parity memories.