Reclaiming free space in a storage system

• US 10,372,334 B2
• Filed: 02/11/2016
• Issued: 08/06/2019
• Est. Priority Date: 02/11/2016
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Hetzler, Steven R., Hineman, Wayne C.
• Primary Examiner(s): Thai, Tuan V
• Application Number:
  US15/042,014
• Publication Number:
  US 20170235672A1
• Time in Patent Office: 1,272 Days
• Field of Search: 711100, 711117, 711118
• CPC Class Codes: G06F 3/0608 Saving storage space on storage systems
• CPC Class Codes: G06F 3/0652 Erasing, e.g. deleting, data cleaning, moving of data to a wastebasket
• CPC Class Codes: G06F 3/067 Distributed or networked storage systems, e.g. storage area networks [SAN], network attached storage [NAS]

Abstract

One embodiment provides a method for reclaiming free space. The method comprises selecting a first blob for reclamation from a first data center; sending a first message to a second data center indicating the first blob is to be reclaimed; sending a second message to the second data center after reclaiming the first blob; receiving a global reclamation complete message from the second data center; reading at least one data set from the first blob; and storing in a write buffer the at least one data set for encoding into a erasure code group in an alternative blob in the first data center. Further, upon receipt of the global reclamation message from the second data center, indicating the first blob is free in the map in the first data center. In one embodiment, selecting the first blob is based on the map indicating free space in the first data center.

Claims

1. A method for reclaiming free space in a storage system, comprising:

selecting a first blob for reclamation from a first data center;

sending a first message to a second data center indicating the first blob is to be reclaimed;

reclaiming the first blob from the first data center;

sending a second message to the second data center after reclaiming the first blob; and

receiving a global reclamation complete message from the second data center;

wherein;

the global reclamation complete message indicates a second blob in the second data center has been reclaimed;

the second blob comprising data associated with data in the first blob;

the storage system comprises a plurality of data centers;

each data center maintains a local erasure code and a map of data stored locally therein; and

the plurality of data centers is protected by a global erasure code distributed across all blobs across all data centers.

2. The method of claim 1, wherein reclaiming comprises:

reading at least one data set from the first blob; and

storing in a write buffer the at least one data set for encoding into an erasure code group in an alternative blob in the first data center.

3. The method of claim 1, further comprising:

upon receipt of the global reclamation message from the second data center, indicating the first blob is free in the map in the first data center.

4. The method of claim 3, further comprising:

sending to the second data center the map indicating free space in the first data center.

5. The method of claim 4, wherein:

selecting the first blob is based on a total amount of free space of all blobs across all data centers.

6. The method of claim 4, wherein:

selecting the first blob is based on a free space rank of all blobs across all data centers.

7. The method of claim 1, wherein:

selecting the first blob is based on the map indicating free space in the first data center.

8. A method for reclaiming free space in a storage system, comprising:

sending a first message to a second data center indicating a first blob is to be reclaimed;

reclaiming the first blob from the first data center;

sending a second message to the second data center after reclaiming the first blob; and

receiving a global reclamation complete message from the second data center, wherein the global reclamation complete message indicates a second blob in the second data center has been reclaimed;

the second blob comprising data associated with data in the first blob.

9. The method of claim 8, wherein reclaiming comprises:

reading at least one data set from the first blob;

storing in a write buffer the at least one data set for encoding into an erasure code group in an alternative blob in the first data center;

upon receipt of the global reclamation message from the second data center, indicating the first blob is free in the map in the first data center; and

sending to the second data center the map indicating free space in the first data center.

10. The method of claim 9, wherein:

selecting the first blob is based on one of;

the map indicating free space in the first data center, a total amount of free space of all blobs across all data centers, or a free space rank of all blobs across all data centers.