Error-handling for write-distribute command in raid mirroring

• US 8,583,853 B1
• Filed: 07/17/2012
• Issued: 11/12/2013
• Est. Priority Date: 05/22/2009
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Lee, Jung-Ik, Grundler, Grant
• Primary Examiner(s): PHAN, TRONG Q
• Application Number:
  US13/551,306
• Time in Patent Office: 483 Days
• Field of Search: 711/100, 711111-114, 711/143, 711154-156, 711161-162, 711/00, 71111-114, 711/43, 71154-156, 71161-162
• US Class Current: 711/100
• CPC Class Codes: G06F 11/1076 Parity data used in redundant arrays of independent storages, e.g. in RAID systems
• CPC Class Codes: G06F 2211/1019 Fast writes, i.e. signaling the host that a write is done before data is written to disk
• CPC Class Codes: G06F 3/0604 Improving or facilitating administration, e.g. storage management
• CPC Class Codes: G06F 3/0659 Command handling arrangements, e.g. command buffers, queues, command scheduling
• CPC Class Codes: G06F 3/0688 Non-volatile semiconductor memory arrays
• CPC Class Codes: G06F 3/0689 Disk arrays, e.g. RAID, JBOD

Abstract

A method includes transmitting, by a processor to an I/O controller, a write-distribute command. The command includes a single copy of data and instructions to write instances of the data to multiple storage locations within a group of storage mediums. For each individual location of the multiple storage locations, the I/O controller transmits a write complete message to the processor after writing an instance of the data to the individual location is complete. If writing the instance of the data to a particular location has an error, the write complete message for the particular individual location is a write error message.

Claims

1. An apparatus comprising:

a processor;

a group of data storage mediums, and

an I/O controller configured to;

receive, from the processor, a write-distribute command that includes a single copy of data and instructions to write instances of the data to multiple storage locations within the group of data storage mediums;

write instances of the data to the multiple storage locations according to the instructions; and

for each individual location of the multiple storage locations;

transmit a write complete message to the processor after writing an instance of the data to the individual location is complete,

wherein the write complete message for a particular individual location is a write error message if writing the instance of the data to the particular location has an error.

2. The apparatus of claim 1, wherein the processor is configured to:

in response to the write error message,

re-transmit the write-distribute command from the processor to the I/O controller.

3. The apparatus of claim 1, wherein the processor is configured to:

in response to the write error message,

transmit a command to the I/O controller to re-write the data to the particular individual location.

4. The apparatus of claim 3, wherein the I/O controller is configured to:

in a response to the command,

copy the data for rewriting at the particular individual location from a cache or buffer.

5. The apparatus of claim 3, wherein the I/O controller is configured to:

in a response to the command,

copy the data for rewriting at the particular individual location from another individual location at which writing the data was error-free.

6. The apparatus of claim 1, wherein the processor is configured to:

if no write complete messages are received at the processor within a timeout period associated with the write-distribute command,

re-transmit the write-distribute command from the processor to the I/O controller.

7. The apparatus of claim 1, wherein the group of data storage mediums includes a plurality of solid-state storage mediums arranged in a RAID array.

8. An apparatus comprising:

a processor;

a group of data storage mediums, and

an I/O controller configured to;

receive, from the processor, a write-distribute command that includes a single copy of data and instructions to write instances of the data to multiple storage locations within the group of data storage mediums;

write instances of the data to the multiple storage locations according to the instructions;

for an individual location of the multiple storage locations,

transmit, from the I/O controller to the processor, a write complete message after writing an instance of the data to the individual location is complete; and

wherein the processor is configured to re-transmit the write-distribute command to the I/O controller, if no write complete messages are received at the processor within a timeout period associated with the write-distribute command.

9. The apparatus of claim 8, wherein the I/O controller is configured to:

transmit from the I/O controller to the processor, a final write completion message after instances of the data have been successfully written to all of the multiple storage locations.

10. The apparatus of claim 8, wherein the I/O controller is configured to:

transmit from the I/O controller to the processor, a final write completion message after instances of the data have been successfully written to all of the multiple storage locations.

11. The apparatus of claim 8, wherein the group of data storage mediums includes a plurality of solid-state storage mediums arranged in a RAID array.

12. A method comprising:

transmitting, from a processor to an I/O controller, a write-distribute command that includes a single copy of data and instructions to write instances of the data to multiple storage locations within a group of storage mediums;

writing instances of the data to the multiple storage locations according to the instructions; and

for each individual location of the multiple storage locations,

transmitting, from the I/O controller to the processor, a write complete message after writing an instance of the data to the individual location is complete,

wherein the write complete message for a particular individual location is a write error message if writing the instance of the data to the particular location has an error.

13. The method of claim 12, further comprising:

in response to the write error message,

re-transmitting the write-distribute command from the processor to the I/O controller.

14. The method of claim 12, further comprising:

in response to the write error message,

transmitting, from the processor to the I/O controller, a command to re-write the data to the particular individual location.

15. The method of claim 14, further comprising,

in a response by the I/O controller to the command,

copying the data for rewriting at the particular individual location from a cache or buffer.

16. The method of claim 14, further comprising,

in a response by the I/O controller to the command,

copying the data for rewriting at the particular individual location from another individual location at which writing the data was error-free.

17. A method comprising:

transmitting, from a processor to an I/O controller, a write-distribute command that includes a single copy of data and instructions to write instances of the data to multiple storage locations within a group of storage mediums;

writing instances of the data to the multiple storage locations according to the instructions; and

for an individual location of the multiple storage locations,

transmitting, from the I/O controller to the processor, a write complete message after writing an instance of the data to the individual location is complete; and

,

if no write complete messages are received at the processor within a timeout period associated with the write-distribute command,

re-transmitting the write-distribute command from the processor to the I/O controller.

18. The method of claim 17, wherein transmitting, from the I/O controller to the processor, a write complete message after writing an instance of the data to the individual location is complete includes:

transmitting a first write completion message after writing an instance of the data to a first of the multiple storage locations is complete.

19. The method of claim 17, wherein transmitting, from the I/O controller to the processor, a write complete message after writing an instance of the data to the individual location is complete includes:

transmitting a first write completion message after writing an instance of the data to a first of the multiple storage locations is complete; and

transmitting a last write completion message after writing an instance of the data to a last of the multiple storage locations is complete.

20. The method of claim 17, wherein transmitting, from the I/O controller to the processor, a write complete message after writing an instance of the data to the individual location is complete includes:

transmitting a final write completion message after instances of the data have been successfully written to all of the multiple storage locations.