System and method for providing space availability notification in a distributed striped volume set

• US 7,797,489 B1
• Filed: 06/01/2007
• Issued: 09/14/2010
• Est. Priority Date: 06/01/2007
• Status: Expired due to Fees

• Current Assignee: NetApp, Inc.
• Original Assignee: NetApp, Inc.
• Inventors: Jernigan, Richard P. IV, Jiang, Tianyu, Hamilton, Eric
• Primary Examiner(s): Bragdon; Reginald G
• Assistant Examiner(s): Talukdar; Arvind
• Application Number:
  US11/756,932
• Time in Patent Office: 1,201 Days
• Field of Search: 711/114
• US Class Current: 711/114
• CPC Class Codes: G06F 3/0617 in relation to availability
• CPC Class Codes: G06F 3/0644 Management of space entities, e.g. partitions, extents, pools
• CPC Class Codes: G06F 3/0689 Disk arrays, e.g. RAID, JBOD

Abstract

A system and method for managing space availability in a distributed striped file system is provided. A master data server is configured to send space availability detection messages to a plurality of data volumes servers hosting constituent volumes of a striped volume set. If one of the constituent volumes in the striped volume set has a low-in-space flag set, then the master data volume instructs all of the constituent volumes to set a low-in-space required flag, and no further writes are accepted for the striped volume set. The low-in-space and low-in-space required flags represent two states, and these states are returned in response to subsequent space availability detection messages from the master data server. A procedure for utilizing reserved space to complete an accepted cross stripe write operation is also provided.

Claims

1. A method for managing space availability in a distributed striped set, comprising:

providing a master data server and configuring the master data server to periodically send a space availability detection message to each constituent volume in a striped volume set;

in response to receiving the space availability detection message at each constituent volume, querying a local file system to determine whether that constituent volume'"'"'s space availability is below a predetermined threshold, indicating a low-in-space state;

returning a response from each constituent volume to said master data server, indicating whether it is in a low-in-space state;

in response to said master data server receiving a message from one or more constituent volumes with a low-in-space state identifier, determining that the striped volume set is in a low-in-space required state; and

sending a message from said master data server to all constituent volumes to set a low-in-space required state identifier.

2. The method as defined in claim 1, further comprising:

in response to said master data server receiving a message from one or more constituent volumes with a low-in-space state identifier, determining that the striped volume set is in a low-in-space required state; and

sending a message from said master data server to all constituent volumes to set a low-in-space required state identifier.

3. The method as defined in claim 1, further comprising:

upon determining at an individual constituent volume that its space availability is not below a predetermined threshold, clearing a low-in-space state identifier.

4. The method as defined in claim 1, further comprising:

receiving responses at said master data server, from said constituent volumes and, if none of said constituent volumes has set a low-in-space state identifier, sending a message to all constituent volumes to clear the low-in-space required state identifier, and begin allowing write requests.

5. The method as defined in claim 1, further comprising:

upon receipt of a space availability detection message at each said constituent volume from said master data server, returning a response that indicates assertion/non-assertion a low-in-space state identifier and a low-in-space required state identifier.

6. The method as defined in claim 1, further comprising:

said predetermined threshold for said space availability including one stripe width plus an amount of reserved space.

7. A method of processing cross stripe requests in a distributed striped file system, comprising:

accepting a cross write request at a first data volume in a striped volume set;

writing the received data to the first volume;

forwarding a set of excess data to a volume striping module of a second data volume;

checking whether the excess data may be written at the second data volume, if a determination is made that user space is not available by simulating an attempt to accommodate an intended write, then determining whether reserved space is available; and

writing, if reserved space is available, the excess data to the reserved space;

notifying the first data volume that the second data volume is in a low-in-space state; and

at said first data volume, setting a low-in-space state identifier.

8. The method as defined in claim 7, further comprising:

at said second volume, if reserved space is not available, denying said write request.

9. A system for managing space availability in a distributed striped file system, comprising:

a plurality of constituent volumes hosted by one or more data volume servers;

a master data server configured to execute a space availability management process whereby a space availability detection message is sent to each data volume server inquiring of the associated data volume'"'"'s space availability, and said data volume servers are further configured to store space availability data structures including a low-in-space flag and a low-in-space required flag; and

said master data server being further configured to accept or deny write access requests for the striped volume set depending upon space availability information.

10. The system as defined in claim 9, further comprising:

said data volume servers are configured to use reserved space if needed to complete a write request in a cross stripe write operation.

11. A computer readable storage medium for managing space availability in a distributed striped volume cluster, the computer readable medium comprising:

program instructions for configuring a master data server to periodically send a space availability inquiry message to each constituent volume in a striped volume set;

program instructions for configuring each data server such that, in response to receiving the space availability inquire message, querying a local file system to determine whether its constituent volume'"'"'s space availability is below a predetermined threshold, indicating a low-in-space state;

program instructions for returning a response to said master data server from each constituent volume indicating whether each constituent volume is in a low-in-space state;

program instructions for returning a response to said master data server from each constituent volume with a low-in-space-state identifier;

determining that the striped volume cluster is in a low-in-space required state; and

program instructions for sending a message from said master data server to all constituent volumes to set a low-in-space required state identifier.

12. A computer readable storage medium for performing a cross stripe write operation in a distributed striped file system, the computer readable medium further comprising:

program instructions for accepting a cross write stripe request at a first data volume in a striped volume set;

program instructions for writing the received data to the first volume;

program instructions for forwarding a set of excess data to a volume striping module of a second data volume;

upon checking whether the excess data may be written at the second data volume, program instructions for determining that user space is not available by simulating an attempt to accommodate an intended write, but determining whether reserved space is available;

program instructions for writing, if reserved space is available, the excess data to the reserved space;

program instructions for notifying the first data volume that the second data volume is in a low-in-space state; and

at said first data volume, program instructions for setting a low-in-space state identifier.

13. The computer readable medium as defined in claim 12, further comprising:

program instructions for configuring a master data server to periodically send a space availability inquiry message to each constituent volume in a striped volume set;

program instructions for configuring each data server such that, in response to receiving the space availability inquire message, querying a local file system to determine whether its constituent volume'"'"'s space availability is below a predetermined threshold, indicating a low-in-space state; and

program instructions for returning a response to said master data server from each constituent volume indicating whether each constituent volume is in a low-in-space state.

14. The computer readable medium as defined in claim 12, further comprising:

at each said constituent volume, upon receipt of a space availability detection message received from said master data server, program instructions for returning a response that indicates assertion/non-assertion of a low-in-space state identifier and a low-in-space required state identifier.

15. The computer readable medium as defined in claim 14, further comprising:

in response to said space availability inquire message, if none of said constituent volumes sends an identifier that it is in a low in space state, then at said master data server, sending a message to all constituent volumes to clear a low in space required indicator.

16. The computer readable medium as defined in claim 15, further comprising:

when none of said constituent volumes is in a low in space state, then allowing further write requests, including cross stripe write requests.