Providing file system quota support for a file system having separated data and metadata

• US 8,892,507 B1
• Filed: 07/09/2013
• Issued: 11/18/2014
• Est. Priority Date: 03/29/2012
• Status: Active Grant

• Current Assignee: Emc IP Holding Company LLC (Dell Technologies Inc.)
• Original Assignee: EMC Corporation (Dell Technologies Inc.)
• Inventors: Dwyer, Thomas J. III, Wijayaratne, Ravi, Muntz, Daniel A., Noveck, David, Veeraswamy, Sairam
• Primary Examiner(s): Ly, Anh
• Application Number:
  US13/937,566
• Time in Patent Office: 497 Days
• Field of Search: 707/783, 707/645, 707/827, 707/E17.005, 707/609, 707/822, 711/163, 711/165, 711/170, 711/112, 711/156
• US Class Current: 707/609
• CPC Class Codes: G06F 16/182 Distributed file systems
• CPC Class Codes: G06F 16/1827 Management specifically adapted to NAS management of storage area networks [SAN] G06F3/067
• CPC Class Codes: G06F 21/6218 to a system of files or objects, e.g. local or distributed file system or database

Abstract

Described are techniques for performing quota management in a distributed file system. An allocation quota specifies an upper limit of storage for allocation for a quota set of the distributed file system. The distributed file system includes metadata file system(s) and data file system(s). For files in the distributed file system, metadata is stored in one of the metadata file systems and file content data is stored in the data file systems. A quota coordinator partitions the allocation quota among the data file systems. The data file systems are allocators of storage and each receives a portion of the allocation quota specifying an upper limit of storage that the data file system has available for allocation for the quota set. Storage is allocated for client requests. The storage allocation is performed by each of the data file systems in accordance with the portion of the allocation quota assigned.

Claims

1. A method for performing quota management in a distributed file system comprising:

receiving, using a processor, an allocation quota specifying an upper limit of storage available for allocation for a quota set of the distributed file system, wherein the distributed file system includes a first set of one or more metadata file systems storing metadata about files in the distributed file system and a second set of data file systems storing file content data for files of the distributed file system, wherein for at least some files in the distributed file system, metadata for each of the at least some files is stored in at least one of the metadata file systems of the first set and content data for each of the at least some files is stored in at least one of the data file systems of the second set;

partitioning, by a coordinator and using a processor, the allocation quota among the data file systems of the second set, wherein each of the data file systems of the second set is an allocator of storage, wherein each of the data file systems of the second set receives a portion of the allocation quota specifying an upper limit of storage that said each data file system has available for allocation for the quota set, wherein the portion of the allocation quota assigned to at least a first of the data file systems includes an amount of reserved storage provisionally granted to the first data file system for servicing storage requests for the quota set;

allocating, using a processor, storage for client requests, wherein said allocating is performed by each of the data file systems of the second set in accordance with the portion of the allocation quota assigned to said each data file system;

receiving, by the coordinator from a second of the data file systems of the second set and using a processor, a storage reservation request for reserved storage provisionally granted to the second data file system for use in servicing storage requests for the quota set;

determining, by the coordinator and using a processor, that there is an insufficient amount of free space of the allocation quota for the quota set to grant the storage reservation request; and

responsive to said determining that there is an insufficient amount of free space, performing, by the coordinator and using a processor, first processing to recall at least some reserved storage which has been provisionally granted to one or more of the data file systems and which has not yet been allocated.

2. The method of claim 1, wherein at a first point in time, the first data file system is assigned a first portion of the allocation quota specifying an upper limit of storage that said first data file system has available for allocation for the quota set, the first portion comprising a first amount of currently allocated storage and a second amount of reserved storage provisionally granted to the first data file system for use in servicing storage requests for the quota set, said first portion having a first size that is a sum of said first amount and said second amount.

3. The method of claim 2, wherein said first processing further comprises:

sending, by the coordinator to each of one or more of the data file systems, a request to recall at least some reserved storage which has been provisionally granted to said each data file system and which has not yet been allocated by said each data file system.

4. The method of claim 3, wherein said first of the data file systems receives, at a second point in time subsequent to the first point in time, the request to recall at least some reserved storage provisionally granted at the first point in time.

5. The method of claim 4, wherein the request specifies a revised size for said first data portion, said revised size being less than said first size thereby indicating that an amount of said allocation quota partitioned by the coordinator for use by the first data file system has been reduced from said first size at said first point in time to said revised size at said second point in time.

6. The method of claim 5, wherein at said first point in time, the second amount of reserved storage has not yet been allocated to service a storage request for the quota set, and wherein between said first point in time and said second point in time, the first data file system has allocated at least some of the second amount of reserved storage since said first point in time thereby leaving a remaining amount of reserved storage space provisionally allocated to said first data file system which has not yet been allocated, said remaining amount being less than said second amount.

7. The method of claim 6, wherein said remaining amount represents a maximum amount of reserved storage of the first data file system available for recall by said coordinator at said second point in time.

8. The method of claim 6, wherein at the second point in time, the first data file system has currently allocated a third amount of storage, said third amount being greater than said revised size for said first data portion specified in the request from the coordinator.

9. The method of claim 8, wherein, responsive to the request, the first data file system sends a response to the coordinator indicating that the request to recall reserved storage has failed to recall any reserved storage from the first data file system since the third amount of storage currently allocated by the first data file system is greater than the revised size.

10. The method of claim 2, further comprising:

determining by the first data file system whether a requested amount of storage for a first client request, if allocated, would exceed the first size of the first portion, said determining including computing a second sum by adding the requested amount to the first amount and comparing the second sum to the first size of the first portion; and

allocating storage for the first client request if the second sum does not exceed the first size of the first portion.

11. The method of claim 10, wherein if the second sum exceeds the first size of the first portion, the first data file system requests additional storage from the coordinator by issuing a second storage reservation request.

12. The method of claim 11, wherein, responsive to the second storage reservation request, the coordinator giants additional storage to the first data file system thereby increasing the first portion from the first size to a larger size, and wherein at any point in time, a total amount of storage representing a sum of storage capacity of the portions of the allocation quota partitioned by the coordinator across all the data file systems of the second set does not exceed the allocation quota.

13. The method of claim 1, wherein the quota set is a portion of a namespace of the distributed file system and wherein the quota set is determined using an attribute of files and directories denoting an owner.

14. The method of claim 1, wherein the quota set is a first quota set that represents objects of the distributed file system for a particular owner indicated by an owner attribute for the objects and a second quota set is a quota tree, said quota tree being a subtree of a tree that represents a namespace of the distributed file system, and wherein two allocation limits are simultaneously applied in connection with determining whether to perform a requested operation resulting in allocating additional storage for an object of the particular owner, a first of the two allocation limits denotes an allocation limit for the subtree and a second of the allocation limits denotes an allocation limit for the particular user, and the allocation quota is the first allocation limit.

15. The method of claim 1, wherein the metadata file system is the coordinator and also an allocator of storage.

16. The method of claim 1, wherein the second set of data file systems includes a plurality of data file systems and file content data for at least a first file of the distributed file system is stored in more than one of the plurality data file systems.

17. The method of claim 16, wherein the file content data for the first file is striped across the more than one of the plurality of data file systems.

18. A computer readable storage medium comprising code stored thereon for performing quota management in a distributed file system, the computer readable storage medium comprising code, that when executed by a processor, performs a method comprising:

receiving an allocation quota specifying an upper limit of storage available for allocation for a quota set of the distributed file system, wherein the distributed file system includes a first set of one or more metadata file systems storing metadata about files in the distributed file system and a second set of data file systems storing file content data for files of the distributed file system, wherein for at least some files in the distributed file system, metadata for each of the at least some files is stored in at least one of the metadata file systems of the first set and content data for each of the at least some files is stored in at least one of the data file systems of the second set;

partitioning, by a coordinator, the allocation quota among the data file systems of the second set, wherein each of the data file systems of the second set is an allocator of storage, wherein each of the data file systems of the second set receives a portion of the allocation quota specifying an upper limit of storage that said each data file system has available for allocation for the quota set, wherein the portion of the allocation quota assigned to at least a first of the data file systems includes an amount of reserved storage provisionally granted to the first data file system for servicing storage requests for the quota set;

allocating storage for client requests, wherein said allocating is performed by each of the data file systems of the second set in accordance with the portion of the allocation quota assigned to said each data file system;

receiving, by the coordinator from a second of the data file systems of the second set, a storage reservation request for reserved storage provisionally granted to the second data file system for use in servicing storage requests for the quota set;

determining, by the coordinator, that there is an insufficient amount of free space of the allocation quota for the quota set to grant the storage reservation request; and

responsive to said determining that there is an insufficient amount of free space, performing, by the coordinator and using a processor, first processing to recall at least some reserved storage which has been provisionally granted to one or more of the data file systems and which has not yet been allocated.

19. The computer readable storage medium of claim 18, wherein at a first point in time, the first data file system is assigned a first portion of the allocation quota specifying an upper limit of storage that said first data file system has available for allocation for the quota set, the first portion comprising a first amount of currently allocated storage and a second amount of reserved storage provisionally granted to the first data file system for use in servicing storage requests for the quota set, said first portion having a first size that is a sum of said first amount and said second amount.

20. A system comprising:

a first server system including a first memory, wherein code for a metadata file system storing metadata about files in a distributed file system is stored in the first memory;

a set of one or more server systems each including a second memory, wherein code for each of a plurality of data file systems is included in the second memory of one of the server systems in the set, said plurality of data file systems storing file content data for files of the distributed file system; and

wherein code for the metadata file system as stored in the first memory further comprises code that, when executed by a processor, performs a method comprising;

receiving an allocation quota specifying an upper limit of storage available for allocation for a quota set of the distributed file system;

partitioning the allocation quota among the plurality of data file systems, wherein each of the plurality of data file systems is an allocator of storage that receives a portion of the allocation quota specifying an upper limit of storage that said each data file system has available for allocation for the quota set, wherein the portion of the allocation quota assigned to at least a first of the plurality of data file systems includes an amount of reserved storage provisionally granted to the first data file system for servicing storage requests for the quota set; and

wherein code for each of the plurality of data file systems stored in the second memory of one of the server systems in the set further comprises code that, when executed by a processor, performs a method comprising;

allocating storage for client requests, wherein said allocating is performed by said each data file system in accordance with the portion of the allocation quota assigned to said each data file system; and

wherein code for the metadata file system as stored in the first memory further comprises code that, when executed by a processor, performs a method comprising;

receiving, from a second of the plurality of data file systems, a storage reservation request for reserved storage provisionally granted to the second data file system for use in servicing storage requests for the quota set;

determining that there is an insufficient amount of free space of the allocation quota for the quota set to grant the storage reservation request; and

responsive to said determining that there is an insufficient amount of free space, performing first processing to recall at least some reserved storage which has been provisionally granted to one or more of the plurality of data file systems and which has not yet been allocated.