Accessing a volume in a distributed environment

• US 8,392,680 B1
• Filed: 03/30/2010
• Issued: 03/05/2013
• Est. Priority Date: 03/30/2010
• Status: Active Grant

• Current Assignee: EMC International Company (Dell Technologies Inc.)
• Original Assignee: EMC International Company (Dell Technologies Inc.)
• Inventors: Natanzon, Assaf, Aharoni, Yuval, Ayzenberg, Lev
• Primary Examiner(s): BATAILLE, PIERRE MICHE
• Application Number:
  US12/749,724
• Time in Patent Office: 1,071 Days
• Field of Search: 711/162, 711/161, 711/E12.058, 711/163, 711/206, 711/203
• US Class Current: 711/162
• CPC Class Codes: G06F 11/1469 Backup restoration techniques
• CPC Class Codes: G06F 11/1471 involving logging of persistent data for recovery
• CPC Class Codes: G06F 11/1474 in transactions G06F16/20 takes precedence
• CPC Class Codes: G06F 11/2023 Failover techniques
• CPC Class Codes: G06F 11/2035 without idle spare hardware
• CPC Class Codes: G06F 11/2069 Management of state, configuration or failover
• CPC Class Codes: G06F 11/2071 using a plurality of controllers
• CPC Class Codes: G06F 2201/815 Virtual middleware or OS functionality using virtual machines to implement generic software techniques for error detection or fault masking G06F11/1484
• CPC Class Codes: G06F 2201/82 Solving problems relating to consistency ensuring consistency in mirrored systems G06F11/2064
• CPC Class Codes: G06F 2201/84 Using snapshots, i.e. a logical point-in-time copy of the data
• CPC Class Codes: G06F 2201/855 Details of asynchronous mirroring using a journal to transfer not-yet-mirrored changes

Abstract

In one aspect, a method includes exposing a set of storage volumes to a host at a requested point in time, in a virtual access mode. The set of storage volumes are handled by distributed virtual consistency groups (CGs) having a background process wherein the distributed virtual CGs update the set of storage volumes to the requested point in time. The method also includes exposing a first service storage volume at a data protection appliance, determining if the virtual CGs have rolled back and using at least one of a central manager and a splitter to account for input/output requests (IOs) when a virtual CG has not rolled back.

Claims

1. A method, comprising:

exposing a set of storage volumes to a host at a requested point in time, in a virtual access mode, the set of storage volumes being handled by distributed virtual consistency groups (CGs), having a background process wherein the distributed virtual CGs update the set of storage volumes to the requested point in time;

exposing a first service storage volume at a data protection appliance (DPA);

determining if the virtual CGs have rolled back; and

using at least one of a central manager and a splitter to account for input/output requests (IOs) when a virtual CG has not rolled back,

wherein using the central manager comprises;

notifying the central manager after a virtual consistency group completes a roll back or a virtual consistency group has less than a predetermined amount of data left to roll back;

waiting for the virtual consistency groups to complete a roll back; and

moving to a trial phase if the virtual consistency groups have completed the roll back.

2. The method of claim 1 wherein moving to the trial phase comprises:

receiving IO data at the DPA;

determining if roll back for all the virtual consistency groups is complete;

sending a new IO to a virtual consistency group if roll back of the virtual consistency groups is not complete;

if roll back for each virtual consistency group is complete;

moving the virtual consistency groups from a virtual mode to a logged access mode;

exposing a second service storage volume at the DPA to new IOs;

failing open IOs on the first service storage volume; and

ceasing exposing the first service storage volume to new IOs.

3. The method of claim 2, further comprising, repeat determining, after a predetermined amount of time, if the roll back for all the virtual consistency groups is complete after a first attempt determines that the roll back for all the virtual consistency groups is not complete.

4. A method, comprising:

exposing a set of storage volumes to a host at a requested point in time, in a virtual access mode, the set of storage volumes being handled by distributed virtual consistency groups (CGs), having a background process wherein the distributed virtual CGs update the set of storage volumes to the requested point in time;

exposing a first service storage volume at a data protection appliance (DPA);

determining if the virtual CGs have rolled back; and

using at least one of a central manager and a splitter to account for input/output requests (IOs) when a virtual CG has not rolled back,

wherein using the splitter comprises;

failing an open IO;

ceasing exposing the first service storage volume at the DPA;

exposing a third service storage volume at the DPA;

at the splitter, receiving notification that the open IO failed on the first service storage volume;

at the splitter, moving to a second serve mode if the first service storage volume cannot be found and the third storage volume is found.

5. The method of claim 4 wherein moving to the second serve mode comprises:

intercepting a write IO to a volume V that is offset by y blocks and has a size x blocks;

sending a write IO having a size that is twice the size of the size x to the third service volume at the DPA; and

writing the data if returned status is a specific vendor unique command;

wherein the write IO of size 2x comprises a first half comprising data written that is new and a second half comprising data that is currently stored at the offset y.

6. The method of claim 4 wherein moving to the second serve mode further comprises splitting the IOs.

7. The method of claim 4 wherein moving to the second serve mode further comprises, at the splitter:

intercepting a read command to the volume V that is offset by y blocks and has a size x bytes;

sending a notification of the read to the third service volume at the DPA sending a write-before-read command to the third service volume at the DPA;

comprising the data in the volume V that is offset by y blocks and has the size x to the DPA; and

sending a read command to at the third service volume of the DPA for the volume V, that is offset by y blocks and has a size x.

8. The method of claim 4, further comprising, if the virtual consistency groups are rolled back:

failing an IO on the third service storage volume; and

exposing a second service storage volume.

9. The method of claim 8, further comprising, at the splitter:

receiving notification that the IO failed at the third storage volume; and

moving to a logged access mode if the third service storage is unavailable, and the second storage is available.

10. An apparatus, comprising:

circuitry to;

expose a set of storage volumes to a host at a requested point in time, in a virtual access mode, the set of storage volumes being handled by distributed virtual consistency groups (CGs), having a background process wherein the distributed virtual CGs update the set of storage volumes to the requested point in time;

expose a first service storage volume at a data protection appliance;

determine if the virtual CGs have rolled back; and

use at least one of a central manager and a splitter to account for IOs when a virtual CG has not rolled back,

wherein the circuitry to use the central manager comprises circuitry to;

notify the central manager after a virtual consistency group completes a roll back or a virtual consistency group has less than a predetermined amount of data left to roll back;

wait for the virtual consistency groups to complete a roll back;

move to a trial phase if the virtual consistency groups have completed the roll back.

11. The apparatus of claim 10 wherein the circuitry comprises at least one of a processor, a memory, programmable logic and logic gates.

12. The apparatus of claim 10 wherein the circuitry to use the central manager comprises circuitry to:

receive IO data at the DPA;

determine if roll back for all the virtual consistency groups is complete;

send a new IO to a virtual consistency group if roll back of the virtual consistency groups is not complete;

if roll back for each virtual consistency group is complete;

move the virtual consistency groups from a virtual mode to a logged access mode;

expose a second service storage volume at the DPA to new IOs;

fail open IOs on the first service storage volume; and

cease exposing the first service storage volume to new IOs;

repeat determining, after a predetermined amount of time, if the roll back for all the virtual consistency groups is complete after a first attempt determines that the roll back for all the virtual consistency groups is not complete.

13. An apparatus, comprising:

circuitry to;

expose a set of storage volumes to a host at a requested point in time, in a virtual access mode, the set of storage volumes being handled by distributed virtual consistency groups (CGs), having a background process wherein the distributed virtual CGs update the set of storage volumes to the requested point in time;

expose a first service storage volume at a data protection appliance;

determine if the virtual CGs have rolled back; and

use at least one of a central manager and a splitter to account for IOs when a virtual CG has not rolled back,

wherein the circuitry to use the splitter comprises circuitry to;

fail an open IO;

cease exposing the first service storage volume at the DPA;

expose a third service storage volume at the DPA;

at the splitter, receive notification that the open IO failed on the first service storage volume;

at the splitter, move to a second serve mode if the first service storage volume cannot be found and the third storage volume is found.

14. The apparatus of claim 13 wherein the instructions to move to the second serve mode comprises instructions to:

intercept a write IO to a volume V that is offset by y blocks and has a size x blocks;

send a write IO having a size that is twice the size of the size x to the third service volume at the DPA; and

write the data if returned status is a specific vendor unique command;

wherein the write IO of size 2x comprises a first half comprising data written that is new and a second half comprising data that is currently stored at the offset y.

15. The apparatus of claim 14 wherein the circuitry to move to the second serve mode further comprises circuitry to, at a splitter:

intercept a read command to the volume V that is offset by y blocks and has a size x bytes;

send a notification of the read to the third service volume at the DPA sending a write-before-read command to the third service volume at the DPA;

comprising the data in the volume V that is offset by y blocks and has the size x to the DPA; and

send a read command to at the third service volume of the DPA for the volume V, that is offset by y blocks and has a size x; and

if the virtual consistency groups are rolled back;

fail an IO on the third service storage volume;

expose a second service storage volume

at the splitter, receive notification that the IO failed at the third storage volume; and

at the splitter, move to a logged access mode if the third service storage is unavailable, and the second storage is available.

16. An article comprising:

a machine-readable medium that stores executable instructions, the instructions causing a machine to;

expose a set of storage volumes to a host at a requested point in time, in a virtual access mode, the set of storage volumes being handled by distributed virtual consistency groups (CGs), having a background process wherein the distributed virtual CGs update the set of storage volumes to the requested point in time;

expose a first service storage volume at a data protection appliance (DPA);

determine if the virtual CGs have rolled back; and

use at least one of a central manager and a splitter to account for input/output requests (IOs) when a virtual CG has not rolled back,

wherein the instructions causing the machine to use the central manager comprises instructions causing the machine to;

notify the central manager after a virtual consistency group completes a roll back or a virtual consistency group has less than a predetermined amount of data left to roll back;

wait for the virtual consistency groups to complete a roll back;

move to a trial phase if the virtual consistency groups have completed the roll back.

17. The article of claim 16 wherein the instructions causing the machine to use the central manager comprises instructions causing the machine to:

receive IO data at the DPA;

determine if roll back for all the virtual consistency groups is complete;

send a new IO to a virtual consistency group if roll back of the virtual consistency groups is not complete;

if roll back for each virtual consistency group is complete;

move the virtual consistency groups from a virtual mode to a logged access mode;

expose a second service storage volume at the DPA to new IOs;

fail open IOs on the first service storage volume; and

cease exposing the first service storage volume to new IOs;

repeat determining, after a predetermined amount of time, if the roll back for all the virtual consistency groups is complete after a first attempt determines that the roll back for all the virtual consistency groups is not complete.

18. An article comprising:

a machine-readable medium that stores executable instructions, the instructions causing a machine to;

expose a set of storage volumes to a host at a requested point in time, in a virtual access mode, the set of storage volumes being handled by distributed virtual consistency groups (CGs), having a background process wherein the distributed virtual CGs update the set of storage volumes to the requested point in time;

expose a first service storage volume at a data protection appliance (DPA);

determine if the virtual CGs have rolled back; and

use at least one of a central manager and a splitter to account for input/output requests (IOs) when a virtual CG has not rolled back,

wherein the instructions causing the machine to use the splitter comprises instructions causing the machine to;

fail an open IO;

cease exposing the first service storage volume at the DPA;

expose a third service storage volume at the DPA;

at the splitter, receive notification that the open IO failed on the first service storage volume;

at the splitter, move to a second serve mode if the first service storage volume cannot be found and the third storage volume is found.

19. The article of claim 18, wherein the instructions to move to the second serve mode comprises instructions to:

intercept a write IO to a volume V that is offset by y blocks and has a size x blocks;

send a write IO having a size that is twice the size of the size x to the third service volume at the DPA; and

write the data if returned status is a specific vendor unique command;

wherein the write IO of size 2x comprises a first half comprising data written that is new and a second half comprising data that is currently stored at the offset y.

20. The article of claim 19 wherein the instructions causing the machine to move to the second serve mode further comprises instructions causing the machine to, at a splitter:

intercept a read command to the volume V that is offset by y blocks and has a size x bytes;

send a notification of the read to the third service volume at the DPA sending a write-before-read command to the third service volume at the DPA;

comprising the data in the volume V that is offset by y blocks and has the size x to the DPA; and

send a read command to at the third service volume of the DPA for the volume V, that is offset by y blocks and has a size x.

21. The article of claim 18, further comprising, instructions causing the machine to, if the virtual consistency groups are rolled back:

fail an IO on the third service storage volume;

expose a second service storage volume

at the splitter, receive notification that the IO failed at the third storage volume; and

at the splitter, move to a logged access mode if the third service storage is unavailable, and the second storage is available.