Redundant storage gateways

• US 8,793,343 B1
• Filed: 08/18/2011
• Issued: 07/29/2014
• Est. Priority Date: 08/18/2011
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Sorenson, James Christopher III, Lin, Yun
• Primary Examiner(s): Avellino, Joseph E
• Assistant Examiner(s): BAILEY, DEFORREST P
• Application Number:
  US13/212,960
• Time in Patent Office: 1,076 Days
• Field of Search: 709/219, 709/201, 709/203
• US Class Current: 709/219
• CPC Class Codes: G06F 11/20 using active fault-masking, e.g. by switching out faulty elements or by switching in spare elements
• CPC Class Codes: G06F 11/2025 using centralised failover control functionality
• CPC Class Codes: G06F 11/2092 Techniques of failing over between control units
• 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 16/00 Information retrieval; Database structures therefor; File system structures therefor
• CPC Class Codes: G06F 16/1827 Management specifically adapted to NAS management of storage area networks [SAN] G06F3/067
• CPC Class Codes: G06F 16/183 Provision of network file services by network file servers, e.g. by using NFS, CIFS network file access protocols H04L67/1097
• CPC Class Codes: H04L 41/0654 using network fault recovery ring fault isolation or reconfiguration in loop networks without recovery actions by a network management system H04L12/437
• CPC Class Codes: H04L 67/1095 Replication or mirroring of data, e.g. scheduling or transport for data synchronisation between network nodes
• 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: H04L 69/40 for recovering from a failure of a protocol instance or entity, e.g. service redundancy protocols, protocol state redundancy or protocol service redirection management of faults, events, alarms or notifications in data switching networks H04L41/06

Abstract

Methods, apparatus, and computer-accessible storage media for providing redundant storage gateways. A client may create a storage gateway group and add storage gateways to the group. The client may assign one or more volumes on a remote data store to each the storage gateways in the group. Volume data for each storage gateway in the group may be replicated to at least one other storage gateway in the group. If one of the gateways in the group becomes unavailable, one or more other gateways in the group may take over volumes previously assigned to the unavailable gateway, using the replicated data in the group to seamlessly resume gateway operations for the respective volumes. Client processes that previously communicated with the unavailable gateway may be manually or automatically directed to the gateway(s) that are taking over the unavailable gateway'"'"'s volumes.

Claims

1. A method, comprising:

configuring two or more storage gateways as a storage gateway group on a local network and assigning one or more volumes to each storage gateway device in the storage gateway group, wherein each storage gateway provides an interface between one or more client processes on the local network and a storage service on a remote network for accessing client data maintained on a remote data store by the storage service;

performing, by each storage gateway in the storage gateway group;

exposing a volume assigned to the storage gateway to the one or more client processes on the local network via one or more I/O ports;

receiving write requests directed to the volume from the one or more client processes via the I/O ports, appending write data indicated by the write requests to a write log corresponding to the volume on a local data store for the storage gateway, recording location information for the write data in the write log to a metadata store corresponding to the volume, and uploading write data from the write log corresponding to the volume to the remote data store;

sending the write data directed to the volume to one or more other storage gateways in the storage gateway group, wherein each of the one or more other storage gateways appends received write data directed to the volume to a write log corresponding to the volume on a local data store for the respective storage gateway; and

receiving write data directed to one or more other volumes from the one or more other storage gateways in the storage gateway group, appending the received write data to one or more other write logs corresponding to the one or more other volumes on the local data store for the storage gateway, and recording location information for the write data in the one or more other write logs to one or more other metadata stores corresponding to the one or more other volumes;

determining that one of the storage gateways in the storage gateway group is unavailable;

selecting another storage gateway in the storage gateway group to take over a different volume that was assigned to the unavailable storage gateway;

the selected storage gateway taking over storage gateway operations for the different volume; and

exposing the different volume to the one or more client processes on the local network via the one or more I/O ports.

2. The method as recited in claim 1, wherein said selecting another one of the storage gateways in the storage gateway group to take over a different volume that was assigned to the unavailable storage gateway is performed by the storage gateway group according to a distributed election technique.

3. The method as recited in claim 1, wherein said taking over storage gateway operations for the different volume comprises:

the selected storage gateway querying one or more other storage gateways in the storage gateway group to determine if any write data is missing from a write log on a local data store for the selected storage gateway, wherein the write log corresponds to the different volume; and

if any write data is missing from the write log on the local data store, the selected storage gateway obtaining the missing write data for the different volume from the one or more other storage gateways in the storage gateway group.

4. The method as recited in claim 1, wherein each storage gateway in the storage gateway group further performs:

sending indications of events related to the volume to the one or more other storage gateways in the storage gateway group, wherein events related to a volume include snapshot events, upload events, and volume ownership events;

receiving indications of events related to the one or more other volumes from the one or more other storage gateways in the storage gateway group; and

recording the events related to the one or more other volumes to the one or more other metadata stores corresponding to the one or more other volumes.

5. A storage gateway, comprising:

at least one processor;

one or more I/O ports; and

a memory comprising program instructions, wherein the program instructions are executable by the at least one processor to;

provide an interface between one or more client processes on a local network and a storage service on a remote network for accessing client data maintained on a remote data store by the storage service;

receive write data directed to a volume from another storage gateway on the local network and store the write data directed to the volume to a local data store, wherein the volume is exposed by the other gateway such that the other gateway uploads the write data to the remote data store for the volume, wherein the storage gateway and the other storage gateway are on the same local network, wherein the local data store is distinct from the storage service on the remote network, wherein the volume is assigned to the other storage gateway; and

upon determining that the other storage gateway is unavailable, take over the volume that was assigned to the other storage gateway, wherein, to take over the volume, the program instructions are executable by the at least one processor to;

begin uploading the write data directed to the volume from the local data store to the remote data store; and

expose, by the storage gateway, the volume to the one or more client processes on the local network via one or more I/O ports such that the volume is assigned to the storage gateway and the storage gateway stores write data for the volume to the local data store and uploads the write data to the remote data store for the volume via the storage service on the remote network.

6. The storage gateway as recited in claim 5, wherein, to store the write data directed to the volume to the local data store, the program instructions are executable by the at least one processor to append the write data received from the other storage gateway to a write log and record write log location information for the write data to a metadata store.

7. The storage gateway as recited in claim 5, wherein the program instructions are further executable by the at least one processor to inform at least one other storage gateway on the local network that the storage gateway has taken over the volume from the unavailable storage gateway.

8. The storage gateway as recited in claim 5, wherein the program instructions are further executable by the at least one processor to, subsequent to said exposing the volume, receive write data directed to the exposed volume from the one or more client processes via the one or more I/O ports and store the write data directed to the exposed volume received from the one or more client processes to the local data store.

9. The storage gateway as recited in claim 8, wherein, to store the write data directed to the exposed volume received from the one or more client processes to the local data store, the program instructions are executable by the at least one processor to append the write data to a write log corresponding to the volume and record write log location information for the write data to a metadata store corresponding to the volume.

10. The storage gateway as recited in claim 8, wherein the program instructions are executable by the at least one processor to send the write data directed to the exposed volume received from the one or more client processes to at least one other storage gateway on the local network.

11. The storage gateway as recited in claim 5, wherein the other storage gateway also sends the write data directed to the volume to at least one other storage gateway on the local network, and wherein, to take over the volume, the program instructions are executable by the at least one processor to:

determine that a portion of the write data for the volume is missing from the local data store; and

obtain the missing portion of the write data for the volume from the at least one other storage gateway.

12. The storage gateway as recited in claim 11, wherein, to determine that a portion of the write data for the volume is missing from the local data store, the storage gateway process is operable to examine sequence numbers for the write data to determine if any sequence numbers are missing from the local data store.

13. The storage gateway as recited in claim 5, wherein the program instructions are further executable by the at least one processor to receive and record indications of events related to the volume from the other storage gateway prior to the other storage gateway becoming unavailable.

14. The storage gateway as recited in claim 13, wherein the events include snapshot events and upload events.

15. The storage gateway as recited in claim 13, wherein the program instructions are further executable by the at least one processor to:

determine that one of the events was not completed by the other storage gateway; and

complete the event related to the volume subsequent to said taking over the volume.

16. The storage gateway as recited in claim 5, wherein the program instructions are further executable by the at least one processor to receive write data directed to one or more other volumes assigned to the storage gateway from the one or more client processes via the one or more I/O ports, store the write data directed to the one or more other volumes to the local data store, and upload the write data directed to the one or more other volumes from the local data store to the remote data store.

17. The storage gateway as recited in claim 16, wherein, to store the write data directed to the one or more other volumes to the local data store, the program instructions are executable by the at least one processor to append the write data to one or more write logs each corresponding to one of the one or more other volumes and record write log location information for the write data to one or more metadata stores each corresponding to one of the one or more other volumes.

18. The storage gateway as recited in claim 17, wherein, to upload the write data directed to the one or more other volumes from the local data store, the program instructions are executable by the at least one processor to asynchronously upload write data from the one or more write logs to the remote data store.

19. The storage gateway as recited in claim 16, wherein the program instructions are further executable by the at least one processor to send the write data directed to the one or more other volumes to at least one other storage gateway on the local network.

20. The storage gateway as recited in claim 16, wherein the program instructions are further executable by the at least one processor to send indications of events related to the one or more other volumes to the at least one other storage gateway on the local network, wherein the events include snapshot events and upload events.

21. A non-transitory computer-accessible storage medium storing program instructions computer-executable to implement a storage gateway on a computing device operable to:

provide an interface between one or more client processes on a local network and a storage service on a remote network for accessing client data maintained on a remote data store by the storage service;

expose one or more volumes to the one or more client processes on the local network via one or more I/O ports, wherein the one or more volumes are assigned to the storage gateway;

receive write data directed to the one or more volumes from the one or more client processes via the one or more I/O ports, store the write data directed to the one or more volumes to a local data store, and upload the write data directed to the one or more volumes from the local data store to the remote data store;

send the write data directed to the one or more volumes to one or more other storage gateways in a storage gateway group on the local network;

receive write data directed to one or more other volumes from the one or more other storage gateways in the storage gateway group and store the write data directed to the one or more other volumes to the local data store, wherein each of the one or more other volumes is assigned to a particular one of the one or more other storage gateways;

wherein the storage gateway is configured to take over at least one volume that was assigned to one of the one or more other storage gateways in the storage gateway group; and

wherein to take over the at least one volume, the storage gateway is configured to, for each volume taken over, expose the volume to the one or more client processes on the local network via the one or more I/O ports such that the volume is assigned to the storage gateway and the storage gateway stores write data for the at least one volume to the local data store and uploads the write data to the remote data store for the at least one volume via the storage service on the remote network.

22. The non-transitory computer-accessible storage medium as recited in claim 21, wherein, to take over the at least one volume, the storage gateway is further operable to, for each volume taken over:

perform a snapshot of write data for the volume from the local data store to the remote data store.

23. The non-transitory computer-accessible storage medium as recited in claim 22, wherein, subsequent to said exposing the volume to the one or more client processes on the local network via the one or more I/O ports, the storage gateway is operable to receive write data directed to the volume from the one or more client processes via the I/O ports, store the write data directed to the volume to the local data store, and upload the stored write data for the volume from the local data store to the remote data store.

24. The non-transitory computer-accessible storage medium as recited in claim 23, wherein, to store the write data directed to the volume to the local data store, the storage gateway is operable to append the write data to a write log corresponding to the volume and record write log location information for the write data to a metadata store corresponding to the volume.

25. The non-transitory computer-accessible storage medium as recited in claim 23, wherein the storage gateway is further operable to send the write data directed to the volume received from the one or more client processes via the I/O ports to the one or more other storage gateways in the storage gateway group.

26. The non-transitory computer-accessible storage medium as recited in claim 21, wherein the storage gateway is further operable to inform at least one other storage gateway on the local network that the storage gateway has taken over the volume from the unavailable storage gateway.

27. The non-transitory computer-accessible storage medium as recited in claim 21, wherein, to store the write data directed to the one or more volumes to the local data store, the storage gateway is operable to append the write data to one or more write logs each corresponding to one of the one or more volumes and record write log location information for the write data to one or more metadata stores each corresponding to one of the one or more volumes, and wherein, to upload the write data directed to the one or more volumes from the local data store to the remote data store, the storage gateway is operable—

to upload the write data from the one or more write logs to the remote data store.

28. The non-transitory computer-accessible storage medium as recited in claim 21, wherein, to store the write data directed to the one or more other volumes to the local data store, the storage gateway is operable to append the write data to one or more write logs each corresponding to at least one of the one or more other volumes and record write log location information for the write data to one or more metadata stores each corresponding to at least one of the one or more other volumes.