Efficient garbage collection in a compressed journal file

• US 8,935,304 B2
• Filed: 03/07/2013
• Issued: 01/13/2015
• Est. Priority Date: 10/17/2011
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Amit, Jonathan, Koifman, Chaim, Marenkov, Sergey, Raikhman, Rostislav
• Primary Examiner(s): Spieler, William
• Application Number:
  US13/789,067
• Publication Number:
  US 20130185338A1
• Time in Patent Office: 677 Days
• Field of Search: 707/813
• US Class Current: 707/813
• CPC Class Codes: G06F 12/0253 Garbage collection, i.e. reclamation of unreferenced memory
• CPC Class Codes: G06F 16/1727 Details of free space management performed by the file system saving storage space on storage systems G06F3/0608; management of blocks in storage devices G06F3/064
• CPC Class Codes: G06F 16/1815 Journaling file systems
• CPC Class Codes: G06F 3/06 Digital input from, or digital output to, record carriers , e.g. RAID, emulated record carriers or networked record carriers
• CPC Class Codes: G06F 3/0608 Saving storage space on storage systems
• CPC Class Codes: G06F 3/064 Management of blocks
• CPC Class Codes: G06F 3/0652 Erasing, e.g. deleting, data cleaning, moving of data to a wastebasket
• CPC Class Codes: G06F 3/0689 Disk arrays, e.g. RAID, JBOD

Abstract

A map corresponding to data blocks with overwritten compressed journal entries is configured. Weighted conditions for each of the overwritten compressed journal entries are calculated. The weighted conditions are arranged in the map from lowest to highest. One of the weighted conditions includes a biasing variable towards selecting data blocks having free space at an end of at least one associated record.

Claims

1. A method for efficient garbage collection in a compressed journal file system by a processor device in a computing environment, the method comprising:

configuring a map corresponding to data blocks with overwritten compressed journal entries; and

calculating weighted conditions for each of the overwritten compressed journal entries, wherein the weighted conditions are arranged in the map from lowest to highest;

wherein one of the weighted conditions includes a biasing variable towards selecting data blocks having free space at an end of at least one associated record.

2. The method of claim 1, further including, performing the calculation for the weighted conditions according to:

(x−

w+(y/x))*(z/range), where w is a number of records in a read cache, x is a number of valid records, y is a size of data in the valid records, z is a relative position of the data block to the end of the at least one record, and range is the range of data blocks with the weighted conditions that are equivalent.

3. The method of claim 1, further including, limiting the size of the map by the age of the data blocks to exclude the data blocks with the highest number of valid records from entering the map, wherein the age of the block is a time difference between the last overwrite operation on the data block and a current time.

4. The method of claim 1, wherein the calculating is performed each time data is overwritten on the data blocks and each time the data blocks are added to the map, wherein the weighted conditions of the data blocks are arranged in the map from lowest to highest.

5. The method of claim 1, further including selecting the data blocks for garbage collection with a minimal number of valid data, minimal valid record count, and maximum size of the free space.

6. The method of claim 5, further including, in conjunction with the selection, determining if the weighted conditions of the data blocks exceed a threshold value.

7. The method of claim 6, further including, performing the selecting for the data blocks present in a read cache with weighted conditions that are equivalent.

8. The method of claim 7, further including, if one of the data blocks is not in the read cache, and if the one of the data blocks weighted conditions exceeds the threshold value, performing the selecting for one of the data blocks.