Request ordering support when switching virtual disk replication logs

• US 9,600,206 B2
• Filed: 08/01/2012
• Issued: 03/21/2017
• Est. Priority Date: 08/01/2012
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Inventors: Kotagiri, Sriravi, Newaskar, Rahul Shrikant
• Primary Examiner(s): PORTKA, GARY J
• Application Number:
  US13/564,449
• Publication Number:
  US 20140040572A1
• Time in Patent Office: 1,693 Days
• Field of Search: 707/999.202, 707/E17.005, 707/E17.007, 711/6, 711/152, 711/154, 711/159, 711/162, 711/163, 711/E12.001, 711/E12.103
• 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 2201/815 Virtual
• CPC Class Codes: G06F 2201/82 Solving problems relating to consistency
• CPC Class Codes: G06F 3/0619 in relation to data integrity, e.g. data losses, bit errors
• CPC Class Codes: G06F 3/065 Replication mechanisms
• CPC Class Codes: G06F 3/0664 at device level, e.g. emulation of a storage device or system
• CPC Class Codes: G06F 3/0676 Magnetic disk device

Abstract

Storage access requests, such as write requests, are received from a virtual machine. A storage request processing module updates one of multiple virtual disks as directed by each of the storage access requests, and a replication management module stores information associated with each storage access request in one of multiple logs. The logs can be transferred to a recovery device at various intervals and/or in response to various events, which results in switching logs so that the replication management module stores the information associated with each storage access request in a new log and the previous (old) log is transferred to the recovery device. During this switching, request ordering for write order dependent requests is maintained at least in part by blocking processing of the information associated with each storage access request.

Claims

1. A method comprising:

receiving storage access requests from a virtual machine in a system in which the storage access requests are received by both a storage request processing module that updates one of multiple virtual disks as directed by each of the storage access requests and a replication management module that stores information associated with each storage access request in one of multiple logs; and

maintaining, when switching the multiple logs, request ordering for write order dependent requests, the switching the multiple logs comprising;

switching the multiple logs in two parts, a first part of the two parts including initializing new logs and blocking processing of the information associated with each storage access request without blocking updating of the multiple virtual disks as directed by the storage access requests.

2. A method as recited in claim 1, each of the multiple virtual disks corresponding to one of the multiple logs.

3. A method as recited in claim 1, the request ordering comprising one storage access request not being issued until confirmation of another storage access request is received in accordance with a write-ahead-logging (WAL) protocol.

4. A method as recited in claim 1, the switching the multiple logs in two parts, the two parts including a second of the two parts including changing to the new logs and unblocking processing of new log queue entries.

5. A method as recited in claim 4, the first part including a first stage in which the new logs are initialized but processing of the information associated with each storage access request is not blocked, and a second stage in which processing of the information associated with each storage access request is blocked, the second stage being performed only after all the new logs are initialized.

6. A method as recited in claim 4, the second part including one stage in which the new logs are changed to and processing of new log queue entries is unblocked, and an additional stage in which switching of the multiple logs is finalized including flushing log queue entries in memory to storage, the processing of new log queue entries being unblocked only after all the new logs are changed to.

7. A method as recited in claim 1, the first part further including unblocking processing of new log queue entries in response to a handle of a service managing switching of the multiple log files being closed.

17. A method as recited in claim 1, further comprising in response to an unexpected request during the switching the multiple logs, undoing changes made as part of the switching including:

deleting new logs that are created; and

preventing old logs from being switched.

18. A method as recited in claim 1, further comprising:

writing the multiple logs to an allocated amount of volatile memory effective to reduce input/output processing; and

writing the multiple logs to a storage device when the allocated amount of the volatile memory reaches a threshold.

8. One or more computer storage media having stored thereon multiple instructions that, when executed by one or more processors of a system, cause the one or more processors to:

receive storage access requests from a virtual machine in the system in which the storage access requests are received by both a storage request processing module and a replication management module;

update, by the storage request processing module for each of the storage access requests, one of multiple virtual disks as directed by the storage access request;

buffer, by the replication management module, each of the storage access requests in parallel with the update performed by the storage request processing module;

store, by the replication management module for each of the buffered storage access requests, information associated with the storage access request in one of multiple logs;

switch the multiple logs and, while switching the multiple logs, blocking processing of the information associated with each storage access request without blocking updating of the multiple virtual disks as directed by the storage access requests, and maintaining request ordering for write order dependent requests; and

transmit, for each of the storage access requests, a response to the virtual machine indicating a respective storage access request as complete when the storage request processing module has completed the update of the one of multiple virtual disks as directed by the respective storage access request and the replication management module has completed the storing of the information associated with the respective storage access request in the one of multiple logs.

9. One or more computer storage media as recited in claim 8, each of the multiple virtual disks corresponding to one of the multiple logs.

10. One or more computer storage media as recited in claim 8, the request ordering comprising one storage access request not being issued until confirmation of another storage access request is received in accordance with a write-ahead-logging (WAL) protocol.

11. One or more computer storage media as recited in claim 8, each of the storage access requests comprising a write request.

12. One or more computer storage media as recited in claim 8, the instructions causing the one or more processors to switch the multiple logs comprising instructions causing the one or more processors to perform switching of the multiple logs in two parts, a first of the two parts including initializing new logs and blocking processing of the information associated with each storage access request, and a second of the two parts including changing to the new logs and unblocking processing of new log queue entries.

13. One or more computer storage media as recited in claim 12, the first part including a first stage in which the new logs are initialized but processing of the information associated with each storage access request is not blocked, and a second stage in which processing of the information associated with each storage access request is blocked, the second stage being performed only after all the new logs are initialized.

14. One or more computer storage media as recited in claim 13, the second part including a third stage in which the new logs are changed to and processing of new log queue entries is unblocked, and a fourth stage in which switching of the multiple logs is finalized including flushing log queue entries in memory to storage, the processing of new log queue entries being unblocked only after all the new logs are changed to.

15. One or more computer storage media as recited in claim 14, the multiple instructions further causing the one or more processors to, in response to an identifier of a service managing switching of the log files being closed, unblock processing of new log queue entries prior to all the new logs being changed to.

19. One or more computer storage media as recited in claim 8, the multiple instructions further causing the one or more processors to:

write the multiple logs to an allocated amount of volatile memory effective to reduce input/output processing; and

write the multiple logs to a storage device when the allocated amount of the volatile memory reaches a threshold.

20. One or more computer storage media as recited in claim 8, the multiple instructions further causing the one or more processors to:

receive, while switching the multiple logs and during the blocking, additional storage access requests from the virtual machine; and

buffer, while switching the multiple logs and during the blocking, each of the additional storage access requests.

16. A system comprising:

a storage request processing module configured to receive storage access requests from a virtual machine in the system in which the storage access requests are received by both the storage request processing module and a replication management module and update, for each of the storage access requests, one of multiple virtual disks as directed by the storage access request;

the replication management module configured to store, for each of the storage access requests, information associated with the storage access request in one of multiple logs, switch the multiple logs in multiple stages, maintain, during the switching, request ordering for write order dependent requests, and transmit, for each of the storage access requests, a response to the virtual machine indicating a respective storage access request as complete when the storage request processing module has completed the update of the one of multiple virtual disks as directed by the respective storage access request and the replication management module has completed the storing of the information associated with the respective storage access request in the one of multiple logs, and the multiple stages including;

a first stage in which new logs are initialized but processing of the information associated with each storage access request is not blocked,

a second stage in which processing of the information associated with each storage access request is blocked without blocking the update of the one of multiple virtual disks as directed by the storage access requests, the second stage being performed only after all the new logs are initialized,

a third stage in which the new logs are changed to and processing of new log queue entries is unblocked, the third stage being performed only after processing of the information associated with each storage access request is blocked, and

a fourth stage in which switching of the multiple logs is finalized including flushing log queue entries in memory to storage, the fourth stage being performed only after processing of the new log queue entries is unblocked.