KR20170002848A - Apparatus and method for journaling based on garbage collection - Google Patents

Abstract

According to one embodiment of the present invention, an apparatus for journaling garbage collection includes: a storage; a memory in which a garbage collection journaling program is stored; and a processor for executing the program. The processor copies a victim block to a free block if garbage collection with respect to at least one victim block occurs upon execution of the program, records an address of the victim block and an address of the free block in a journal area, and stores the free block in the storage after the copying to the free block is completed and stores the journal area in a storage device when the storing of the free block is completed. The storage includes a file system, and the file system includes a victim block, a free block, and a journal area.

Description

[0001] APPARATUS AND METHOD FOR JOURNALING BASED ON GARBAGE COLLECTION [0002]

The present invention relates to a garbage collection journaling apparatus and method.

A log structured file system processes user write requests sequentially to storage. The log-structured file system is suitable for NAND flash storage that does not overwrite data according to these characteristics. The log-structured file system is an out-of-the-box update that is a way to invalidate data written to a block where originally modified data is stored and to write to another block in which data is not stored, (out-of-update). Therefore, the log-structured file system can always process random writes to the storage requested by the user in sequence.

In most storage, sequential write has higher write performance than random write. In particular, sequential writing in NAND flash storage with a fixed number of writes, such as a solid state drive, can improve the write speed and extend the life of NAND flash storage.

However, in order to perform continuous sequential writes, the log-structured file system must perform garbage collection. In garbage collection, a valid block is copied to another block in a space where a valid block and an invalid block are mixed, thereby obtaining a usable block that can be overwritten. However, since sequential writes may not be able to handle user write requests during garbage collection, write latency may occur, which may reduce write performance. As such, the log structure file system can reduce write performance due to garbage collection, and may incur additional costs.

Also, even if garbage collection is not performed, a block in which data is stored may be changed from time to time. Therefore, the log structure file system must modify and store metadata indicating the data. Therefore, the log structure file system may incur additional costs for metadata modification. To solve this problem, the log structure file system can use the metadata modification delay technique. However, the metadata modification delay technique can cause garbage collection to break the consistency between storage and file systems.

Conventional techniques for solving such problems include file systems such as segment space recycling-log structured file system (SSR-LFS) and flash friendly file system (F2FS). SSR-LFS reduces the additional write requests by keeping indirect blocks, which are frequently modified metadata, in the cache of main memory based on the lazy indirect block update (LIBU) technique. Since SSR-LFS maintains metadata such as segment information in a storage device, it can solve the problem of inconsistency in garbage collection. However, since SSR-LFS does not keep all metadata in main memory, it can not completely solve the problem of writing metadata in a log-structured file system.

F2FS is a log-structured file system that uses modification delay techniques for all metadata. F2FS uses checkpoints to resolve consistency problems when performing garbage collection. Checkpoints can store all the modified metadata of the file system in storage after garbage collection is done. Therefore, F2FS can solve the consistency problem. However, because F2FS has a large number of modified metadata that is not written to storage, checkpoint costs can increase. And when F2FS performs a checkpoint when performing garbage collection, modified metadata that is not related to garbage collection can be recorded. Therefore, F2FS can increase garbage collection time and may degrade write performance.

In order to solve such a problem, Korean Patent Laid-Open Publication No. 10-2007-0061345 (entitled "Method for securing data storage space using double journaling") discloses a method for securing a data storage space using double journaling in a flash memory . Specifically, the method divides the free list of the segment for the flash memory into a first half journaling free list and a second half journaling free list by setting a center point, and judges whether a free segment is sufficiently secured from the first half free list . In this method, when the free segment is not sufficiently secured, garbage collection is performed by referring to the threshold value of the erasure leveling and the limit value of the effective data set in the segment in which the predetermined data is already stored, thereby securing the data storage space.

In addition, Korean Patent Laid-Open No. 10-2008-0016058 (entitled " Java virtual machine memory management device and method ") discloses a memory management method of a Java virtual machine Apparatus and method.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a garbage collection journaling apparatus and method for performing garbage collection in a log structure file system.

It should be understood, however, that the technical scope of the present invention is not limited to the above-described technical problems, and other technical problems may exist.

According to a first aspect of the present invention, there is provided a garbage collection journaling apparatus including a storage unit, a memory for storing a garbage collection journaling program, and a processor for executing the program, When garbage collection for one or more victim blocks occurs, the victim block is copied to the free block, the address of the victim block and the address of the free block are written to the journal area, and after the copying to the free block is completed, And stores the journal area in the storage device when the storage of the free block is completed. At this time, the storage includes a file system, and the file system includes a sacrifice block, a free block, and a journal area.

A method for garbage collection of a computing device according to a second aspect of the present invention includes the steps of: (a) copying a victim block into a free block as garbage collection for one or more victim blocks occurs; (b) writing the address of the victim block and the address of the free block in the journal area; (c) after step (b), storing the free blocks in the storage; And (d) after step (c), storing the journal area in a storage device. At this time, the storage includes a file system, and the file system includes a sacrifice block, a free block, and a journal area.

According to any one of the above-mentioned problems, the present invention can ensure consistency of data by performing garbage collection instead of checkpoint when performing garbage collection in a log structure file system applying the metadata modification delay technique . Unlike the conventional checkpoint for recording all the modified metadata of the file system, the present invention records only a few selected journal blocks included in the metadata, thereby reducing writing costs in performing garbage collection. Therefore, the present invention can reduce the waiting time of the user's write request that can be delayed while garbage collection is performed, and the write performance to the storage can be increased.

Figure 1 is an example of a consistency problem between storage and file systems.
2 is a block diagram of a garbage collection journaling apparatus according to an embodiment of the present invention.
3 is an illustration of garbage collection according to an embodiment of the present invention.
4 is an exemplary view of a journal area and a journal block of a file system according to an embodiment of the present invention.
5 is a flowchart of a garbage collection journaling method of a garbage collection journaling apparatus according to an embodiment of the present invention.
6 is a flowchart of a method of recovering a file system of a garbage collection journaling apparatus according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings, which will be readily apparent to those skilled in the art. The present invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. In order to clearly explain the present invention in the drawings, parts not related to the description are omitted.

Throughout the specification, when a part is referred to as being "connected" to another part, it includes not only "directly connected" but also "electrically connected" with another part in between . Also, when a part is referred to as "including " an element, it does not exclude other elements unless specifically stated otherwise.

Conventional log-structured file systems that use the metadata modification delay technique may suffer from a break in consistency between storage and file systems when performing garbage collection. The consistency problem will be described with reference to FIG.

Figure 1 is an illustration of a consistency problem between storage and conventional logflex file systems.

FIG. 1 (a) is an example of a checkpoint state before garbage collection is performed. 'File A' and 'File B' are allocated to 'Segment 11' included in the conventional log structure file system. At this time, 'segment 11' is a mixture of valid blocks and invalid blocks. In Fig. 1 (a), there is no block to which a file is allocated in 'segment 27'. A segment made up of only invalid blocks that do not have a block allocated to such a file is called a free segment.

As shown in FIG. 1B, the conventional log structure file system can perform garbage collection for moving blocks allocated to 'file A' and 'file B', respectively, to 'segment 27'. In the conventional log structure file system, blocks assigned to 'file A' and 'file B' can be set as invalid blocks. Therefore, the conventional log structure file system can allocate the blocks 'c1', 'c2' and 'c3' included in the new 'file C' to the first to third blocks in the 'segment 11'.

At this time, if restoration is requested in a state where 'c1', 'c2' and 'c3' are allocated to the block of 'segment 11' before the metadata moved to 'segment 27' is stored, , The conventional log structure file system may attempt to restore 'file A' and 'file B' to the previously allocated 'segment 11'. At this time, since the conventional log structure file system restores to a block of 'segment 11' previously allocated to 'file A' and 'file B' instead of 'segment 27', the recovery fails and the consistency is broken.

Next, a garbage collection journaling apparatus 200 according to an embodiment of the present invention will be described with reference to FIG. 2 to FIG.

2 is a block diagram of a garbage collection journaling apparatus 200 according to an embodiment of the present invention.

The garbage collection journaling apparatus 200 supports a file system for performing garbage collection journaling. At this time, the file system may be executed by an operating system included in the garbage collection journaling apparatus 200. [ For example, the file system may be a log-structured file system.

In addition, the garbage collection journaling device 200 includes a memory 210, a storage 220, and a processor 230.

The memory 210 stores a garbage collection journaling program. At this time, the memory 210 collectively refers to a non-volatile storage device that keeps stored information even when power is not supplied, and a volatile storage device that requires power to maintain stored information. The garbage collection journaling program stored in the memory 210 may be system software or application software executing in the operating system.

The storage 220 may store data. The storage 220 may be embedded in the garbage collection journaling apparatus 200 as a storage device such as a solid state driver and a hard disk driver. The storage 220 may also be a storage device connected to the garbage collection journaling device 200 via a computing bus module. In addition, the storage 220 may be a storage device or a storage server connected through a communication module (not shown) included in the garbage collection journaling device 200.

Upon execution of the garbage collection journaling program stored in the memory 210, the processor 230 copies the victim block to the free block when garbage collection for the victim block stored in the file system occurs. At this time, the file system includes one or more segments including one or more blocks and a journal area. In addition, the file system may include one or more victim blocks when garbage collection occurs.

Specifically, the victim block is a valid block, which is a target of garbage collection. That is, the victim block may be one or more blocks arbitrarily allocated to the file, not sequential to the file system. The victim block is included in the victim segment that will perform garbage collection in the file system. Then, the victim block is moved to the free block, which is an invalid block of the free segment included in the file system. In this case, if there are a plurality of sacrificial blocks, the block to which the sacrificial block is to be moved may be a free block capable of sequentially moving data.

The sacrificial block and free block may be included in the same segment and may be located in different segments, respectively. Also, the sacrificial block and free block may be located in one or more segments.

Then, the processor 230 moves the victim block to the free block, and then writes the addresses of the victim block and the free block in the journal area of the file system. At this time, the journal area of the file system may include one or more journal blocks. For example, the processor 230 may write a pair of addresses of a victim block and a free block in a journal block included in a journal area of the file system. The processor 230 moves all of the victim block to the free block, and then stores the free block in which the victim block is moved in the storage 220. [ When the storage of the free block in which the sacrificial block is moved to the storage 220 is completed, the processor 230 stores the journal area of the file system in the storage 220. The processor 230 may set the victim block into an invalid block so that the journal area of the file system is stored in the storage 220 and the new data can be stored in the victim block.

In this manner, the processor 230 stores the free area in which the sacrificial block is copied in the storage 200, stores the journal area of the file system in the storage 220, and stores the journal area of the file system in the storage after the garbage collection The new data can be stored in the block in which < RTI ID = 0.0 > Therefore, the processor 230 can guarantee the consistency of the data even if the file system is recovered due to a problem in the file system due to sudden power failure or power shutdown during the execution of the garbage collection.

A specific procedure of performing garbage collection will be described with reference to FIG.

3 is an illustration of garbage collection according to an embodiment of the present invention.

First, in order to perform garbage collection, the processor 230 divides the blocks 'B1' to 'B3', which are respectively stored in the sacrifice blocks '3072', '3073' (P300) sequentially to the invalid blocks '1536' to '1538'. In addition, the processor 230 may store the existing allocation information of the victim block and the allocation information after the garbage collection in the journal area of the file system, as shown in FIG.

The processor 230 may then store each allocated block in the main area of the storage 220 (P310). The processor 230 may store the journal area of the file system of the storage 220 in the meta area of the storage 220 (P320), if the block is stored in the main area of the storage 220. [ For example, the processor 230 may store the journal area of the file system in the journal area included in the meta area of the storage 220. [

The processor 230 may then set the victim block of the file system to an invalid block after the journal area of the file system is stored in the meta area of the storage 220. [ Then, the processor 230 sets the sacrifice segment including the sacrifice block as a free segment.

In addition, the processor 230 may store the address of the victim block and the free block in the journal block included in the journal area of the file system, and then write the checkpoint version to the journal block storing the addresses of the victim block and the free block. have. At this time, the checkpoint version is changed every time a checkpoint occurs. That is, the version of the checkpoint stored in the journal area of the file system may be the version of the most recent checkpoint. Therefore, the processor 230 can check when the addresses of the victim block and the free block are stored and whether or not the checkpoint is performed through the checkpoint version stored in the journal area of the file system.

On the other hand, the processor 230 can perform a checkpoint when there is insufficient space to store in the journal area of the file system. The processor 230 may also perform a checkpoint at the request of the operating system or user. And when the checkpoint is performed, the processor 230 may update the version of the checkpoint.

When a plurality of journal blocks are included in the journal area of the file system, the processor 230 may store the addresses of the victim block and the free block and the checkpoint version in the journal block sequentially from the first located journal block. At this time, the journal block may include addresses of one or more of the victim block and the free block. In addition, when a checkpoint occurs, the processor 230 may record the addresses of the victim block and the free block in the journal block sequentially from the first journal block among the plurality of journal blocks. The journal area and journal block of the specific file system will be described with reference to FIG.

4 is an exemplary view of a journal area and a journal block of a file system according to an embodiment of the present invention.

In FIG. 4, the journal area of the file system may include journal block 1 to journal block 3. At this time, when the process executes the checkpoint in the checkpoint version 3, the checkpoint version can be updated from 3 to 4.

After the checkpoint is performed, when the garbage collection is performed, the processor 230 reads the address pair of the victim block and the free block, which performs garbage collection in 'Journal Block 0', which is the first journal block included in the journal area of the file system, Lt; / RTI > In addition, 'checkpoint version 4' which is the current checkpoint version can be recorded in 'journal block 0'. When the garbage collection is performed again, the processor 230 stores the address pair of the victim block and the free block in 'journal block 1' located after 'journal block 0', and updates the checkpoint version 4 ' Can be recorded.

Therefore, when confirming the journal area of the file system, 'journal block 0' and 'journal block 1' storing the latest checkpoint version are recorded again after the checkpoint is performed, and 'journal block 2' It can be confirmed that it is recorded after the point execution.

On the other hand, when the recovery request for the file system occurs, the processor 230 recovers the journal block in which the checkpoint version identical to the checkpoint version recorded at the end of one or more journal blocks included in the journal area of the file system is recorded . That is, the processor 230 may recover the file system based on the journal block that was written after the most recent checkpoint occurred.

For example, referring back to FIG. 4, when a recovery request for a file system occurs, the processor 230 can check the checkpoint version of three journal blocks included in the journal area of the file system. At this time, if the last stored checkpoint version is 'checkpoint version 4', the processor 230 can select 'journal block 0' and 'journal block 1' having checkpoint version 'checkpoint version 4' have. The processor 230 may then perform recovery based on the selected journal block.

Next, a garbage collection journaling method of the garbage collection journaling apparatus 200 will be described with reference to FIGS. 5 and 6. FIG.

5 is a flowchart of a garbage collection journaling method of the garbage collection journaling apparatus 200 according to an embodiment of the present invention.

When the garbage collection for one or more sacrificial blocks occurs, the garbage collection journaling device 200 copies the sacrificial blocks to the free blocks (S500). At this time, the victim block is a block to which the user's file is allocated, and may be a block that is not sequentially allocated. And the free block may mean an invalid block that can sequentially allocate a victim block. In addition, the victim block and free block may be included in the file system of the storage 220.

The garbage collection journaling apparatus 200 records the address of the victim block and the address of the free block in the journal area of the file system (S510). Then, the garbage collection journaling apparatus 200 may write the address of the victim block and the address of the free block in the journal area of the file system, and then write the check point version to the journal area of the file system (S520). At this time, the journal area of the file system may include one or more journal blocks. The journal block can record the address of the victim block and the address of the free block in pairs. Further, the journal block can record the checkpoint version after storing the address of the victim block and the address of the free block.

After the garbage collection journaling apparatus 200 records the addresses of the victim block and the free block in the journal area of the file system, the free block in which the victim block is copied is stored in the storage 220 (S540).

The garbage collection journaling apparatus 200 stores the copied free block in the storage 220, and then stores the journal area of the file system in the storage 220 (S530). The garbage collection journal apparatus 200 may store new data in the victim block in which the garbage collection has been performed since the journal area of the file system is stored in the storage 220.

Through the above process, the garbage collection journaling apparatus 200 according to an embodiment of the present invention stores the free block in which the victim block is copied in the storage 200, stores the journal area of the file system in the storage, After the journal area is stored in the storage 220, new data can be stored in the block in which garbage collection has been performed. Therefore, the garbage collection journaling apparatus 200 can guarantee the consistency of data even if a problem occurs in the file system due to abrupt power failure or power shutdown during garbage collection.

6 is a flowchart of a file system recovery method of a garbage collection journaling apparatus 200 according to an embodiment of the present invention.

If a recovery request for the file system occurs (S600), the garbage collection journaling device 200 checks the same checkpoint as the version of the last recorded checkpoint in the file system for one or more journal blocks included in the journal area of the file system Version of the journal block.

Then, the garbage collection journaling apparatus 200 can perform restoration based on the selected journal block (S610).

The garbage collection journaling apparatus 200 and method according to an embodiment of the present invention perform garbage collection instead of a check point when performing garbage collection in a log structure file system applying the metadata modification delay technique, . Unlike the conventional checkpoint for recording all the modified metadata of the file system, the garbage collection journaling apparatus 200 and method record only a few selected journal blocks included in the metadata, Can be reduced. Thus, the garbage collection journaling device 200 and method may reduce the latency of a user's write request that may be delayed while garbage collection is performed, thereby increasing write performance for the storage 220. [

One embodiment of the present invention may also be embodied in the form of a recording medium including instructions executable by a computer, such as program modules, being executed by a computer. Computer readable media can be any available media that can be accessed by a computer and includes both volatile and nonvolatile media, removable and non-removable media. In addition, the computer-readable medium can include both computer storage media and communication media. Computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Communication media typically includes any information delivery media, including computer readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave, or other transport mechanism.

While the methods and systems of the present invention have been described in connection with specific embodiments, some or all of those elements or operations may be implemented using a computer system having a general purpose hardware architecture.

It will be understood by those skilled in the art that the foregoing description of the present invention is for illustrative purposes only and that those of ordinary skill in the art can readily understand that various changes and modifications may be made without departing from the spirit or essential characteristics of the present invention. will be. It is therefore to be understood that the above-described embodiments are illustrative in all aspects and not restrictive. For example, each component described as a single entity may be distributed and implemented, and components described as being distributed may also be implemented in a combined form.

The scope of the present invention is defined by the appended claims rather than the detailed description and all changes or modifications derived from the meaning and scope of the claims and their equivalents are to be construed as being included within the scope of the present invention do.

200: Garbage collection journaling device
210: memory
220: Storage
230: Processor

Claims (10)

A garbage collection journaling apparatus,
storage,
The memory in which the garbage collection journaling program is stored
And a processor for executing the program,
Wherein, when the garbage collection for one or more victim blocks occurs, the processor copies the victim block to a free block,
The address of the victim block and the address of the free block in the journal area,
Wherein the controller is configured to store the copied free block in the storage after the copying to the free block is completed and store the journal area in the storage when the copying of the completed free block is completed,
Wherein the storage comprises a file system,
Wherein the file system comprises the victim block, the free block and the journal area. The method according to claim 1,
Wherein the processor records the address of the victim block and the address of the free block in the journal area and then writes the checkpoint version in the journal area. The method according to claim 1,
Wherein the processor performs a checkpoint, updates a checkpoint version, and writes the updated checkpoint version to the file system if there is insufficient space to write to the journal area. The method according to claim 1,
Wherein the journal area comprises one or more journal blocks. 5. The method of claim 4,
If garbage collection occurs after performing the checkpoint,
Wherein the address of the victim block and the address of the free block are sequentially recorded sequentially from a journal block located at the beginning of the one or more journal blocks. 5. The method of claim 4,
Wherein the processor recovers, when a recovery request for the file system occurs, a journal block in which a checkpoint version identical to a checkpoint version lastly recorded in the file system of the one or more journal blocks is recorded. A garbage collection journaling method of a garbage collection journaling apparatus,
(a) copying the victim block to a free block as garbage collection for one or more victim blocks occurs;
(b) recording an address of the victim block and an address of the free block in a journal area;
(c) after the step (b), storing the copied free block in the storage; And
(d) after the step (c), storing the journal area in the storage,
Wherein the storage comprises a file system,
Wherein the file system comprises the victim block, the free block and the journal area. 8. The method of claim 7,
After the step (b), before the step (c)
(e) recording a checkpoint version in the journal area. 8. The method of claim 7,
(f) when a recovery request for the file system occurs, for a journal block included in the journal area, a journal block in which a checkpoint version identical to a version of a checkpoint lastly recorded in the file system is recorded, The garbage collection method further comprising: A computer-readable recording medium recording a program for performing the method according to any one of claims 7 to 9 on a computer.

Images