Handling copy on first write data for snapshot purposes

• US 8,874,524 B1
• Filed: 06/22/2012
• Issued: 10/28/2014
• Est. Priority Date: 06/22/2012
• Status: Active Grant

• Current Assignee: Emc IP Holding Company LLC (Dell Technologies Inc.)
• Original Assignee: EMC Corporation (Dell Technologies Inc.)
• Inventors: Zhao, Junping, Xu, Kevin, Xie, Gang, Yang, Jialin, Qi, Wei, Bassov, Ivan
• Primary Examiner(s): Ly, Cheyne D
• Application Number:
  US13/530,503
• Time in Patent Office: 858 Days
• Field of Search: 707/639, 707/649, 707/657, 711/118, 711/162
• US Class Current: 707/657
• CPC Class Codes: G06F 11/2071 using a plurality of controllers
• CPC Class Codes: G06F 16/128 Details of file system snapshots on the file-level, e.g. snapshot creation, administration, deletion use of snapshots for error detection or correction G06F11/14, G06F11/16
• CPC Class Codes: G06F 2201/84 Using snapshots, i.e. a logical point-in-time copy of the data

Abstract

A technique performs a write operation. The technique involves receiving, after a snapshot of a file system is taken, a write instruction to write new data to a particular portion of the file system. The technique further involves includes reading, in response to the write instruction, metadata to determine whether the new data is a first update of the particular portion since the snapshot was taken. The metadata identifies portions of the file system which have changed since the snapshot was taken. The technique further involves, when the new data is a first update of the particular portion since the snapshot was taken, copying the original data from cache memory (i.e., from the file system buffer cache) to snapshot storage (i.e., to a dedicated save area) and, after the original data is copied, updating the original data in the cache memory with the new data.

Claims

1. In an electronic apparatus having cache memory, a method of performing a write operation, the method comprising:

after a snapshot of a file system is taken, receiving a write instruction to write new data to a particular portion of the file system;

in response to the write instruction, reading metadata to determine whether the new data is a first update of the particular portion of the file system since the snapshot was taken, the metadata being constructed and arranged to identify portions of the file system which have changed since the snapshot was taken;

when the metadata indicates that the new data is not a first update of the particular portion of the file system since the snapshot was taken, updating original data in the cache memory with the new data, the original data corresponding to the particular portion of the file system; and

when the metadata indicates that the new data is a first update of the particular portion of the file system since the snapshot was taken, copying the original data from the cache memory to snapshot storage which is separate from the cache memory and, after the original data is copied from the cache memory to the snapshot storage, updating the original data in the cache memory with the new data;

wherein the portions of the file system are data blocks;

wherein the metadata includes a first update table having a set of first update entries, each first update entry corresponding to a respective data block of the file system;

wherein the write instruction identifies a particular data block; and

wherein reading the metadata to determine whether the new data is a first update of the particular portion of the file system since the snapshot was taken includes reading, from the first update table, contents of a particular first update entry corresponding to the particular data block identified by the write instruction, the new data (i) not being a first update of the particular portion of the file system since the snapshot was taken when the contents of the particular first update entry define a first value, and (ii) being a first update of the particular portion of the file system since the snapshot was taken when the contents of the particular first update entry define a second value which is different than the first value.

2. A method as in claim 1 wherein the metadata further includes a pointer table having a set of pointer entries, each pointer entry corresponding to a respective data block of the file system;

and wherein the method further comprises;

in response to copying the original data from the cache memory to the snapshot storage, storing a pointer in a particular pointer entry corresponding to the particular data block identified by the write instruction, the pointer being an address of the snapshot storage which stores the original data copied from the cache memory.

3. A method as in claim 2, further comprising:

after the snapshot of the file system is taken and prior to receiving the write instruction, receiving a read instruction which identifies the particular data block; and

copying the original data from disk storage into the cache memory in response to the read instruction.

4. A method as in claim 2 wherein the write instruction identifies a read-modify-write operation;

and wherein the method further comprises;

copying the original data from disk storage into the cache memory in response to the write instruction.

5. A method as in claim 2 wherein copying the original data from the cache memory to snapshot storage includes:

refraining from modifying the original data in the cache memory while the original data is copied to the snapshot storage and enabling modification of the original data in the cache memory only after the original data is fully copied to the snapshot storage.

6. A method as in claim 2 wherein the electronic apparatus having the cache memory is electronically coupled to another apparatus which is constructed and arranged to mirror the file system;

and wherein the method further comprises;

directing the other electronic apparatus to process the write instruction to replicate the new data at the other electronic apparatus.

7. A method as in claim 2, further comprising:

receiving a snapshot command, and

in response to the snapshot command and prior to receiving the write instruction, creating (i) a snapshot bitmap as the first update table, (ii) a snapshot blockmap as the pointer table, and (iii) a SavVol memory structure as the snapshot storage.

8. A method as in claim 7 wherein copying the original data from the cache memory to the snapshot storage includes:

reading the original data from the cache memory and storing the original data as Copy on First Write (CoFW) data in the SavVol memory structure.

9. A method as in claim 8 wherein reading the original data from the cache memory includes:

performing an efficient hashing operation to locate the CoFW data within the cache memory to avoid re-reading the CoFW data from disk storage.

10. A method as in claim 8 wherein the file system represents a source volume at the time the snapshot of the file system is taken;

and wherein the method further comprises;

after the CoFW data is stored in the SavVol memory structure, receiving a roll back command; and

in response to the roll back command, copying the CoFW data from the SavVol memory structure to the cache memory to restore the source volume.

11. A method as in claim 8 wherein the electronic apparatus is a storage server;

wherein the write instruction originates from a host device which stores data into and loads data from the storage server through a network; and

wherein the snapshot command and the roll back command originate from an administration device which is separate from the electronic apparatus and the host device.

12. An electronic apparatus, comprising:

cache memory;

snapshot storage which is separate from the cache memory; and

a control circuit coupled to the cache memory and the snapshot storage, the control circuit being constructed and arranged to;

after a snapshot of a file system is taken, receive a write instruction to write new data to a particular portion of the file system,

in response to the write instruction, read metadata to determine whether the new data is a first update of the particular portion of the file system since the snapshot was taken, the metadata being constructed and arranged to identify portions of the file system which have changed since the snapshot was taken,

when the metadata indicates that the new data is not a first update of the particular portion of the file system since the snapshot was taken, update original data in the cache memory with the new data, the original data corresponding to the particular portion of the file system, and

when the metadata indicates that the new data is a first update of the particular portion of the file system since the snapshot was taken, copy the original data from the cache memory to the snapshot storage and, after the original data is copied from the cache memory to the snapshot storage, update the original data in the cache memory with the new data;

wherein the portions of the file system are data blocks;

wherein the metadata includes a snapshot bitmap having a set of bitmap entries, each bitmap entry corresponding to a respective data block of the file system;

wherein the write instruction identifies a particular data block; and

wherein the control circuit, when reading the metadata to determine whether the new data is a first update of the particular portion of the file system since the snapshot was taken, is constructed and arranged to read, from the snapshot bitmap, contents of a particular bitmap entry corresponding to the particular data block identified by the write instruction, the new data (i) not being a first update of the particular portion of the file system since the snapshot was taken when the contents of the particular bitmap entry define a first value, and (ii) being a first update of the particular portion of the file system since the snapshot was taken when the contents of the particular bitmap entry define a second value which is different than the first value.

13. An electronic apparatus as in claim 12 wherein the metadata further includes a snapshot blockmap having a set of blockmap entries, each blockmap entry corresponding to a respective data block of the file system;

and wherein the control circuit is further constructed and arranged to;

in response to copying the original data from the cache memory to the snapshot storage, store a pointer in a particular blockmap entry corresponding to the particular data block identified by the write instruction, the pointer being an address of the snapshot storage which stores the original data copied from the cache memory.

14. A computer program product which includes a non-transitory computer readable medium storing containing code to perform a write operation, the code directing a computerized device to:

receive, after a snapshot of a file system is taken, a write instruction to write new data to a particular portion of the file system;

in response to the write instruction, read metadata to determine whether the new data is a first update of the particular portion of the file system since the snapshot was taken, the metadata being constructed and arranged to identify portions of the file system which have changed since the snapshot was taken;

when the metadata indicates that the new data is not a first update of the particular portion of the file system since the snapshot was taken, update original data in cache memory with the new data, the original data corresponding to the particular portion of the file system; and

when the metadata indicates that the new data is a first update of the particular portion of the file system since the snapshot was taken, copy the original data from the cache memory to snapshot storage and, after the original data is copied from the cache memory to the snapshot storage, update the original data in the cache memory with the new data;

wherein the portions of the file system are data blocks;

wherein the metadata includes a snapshot bitmap having a set of bitmap entries, each bitmap entry corresponding to a respective data block of the file system;

wherein the write instruction identifies a particular data block; and

wherein reading the metadata to determine whether the new data is a first update of the particular portion of the file system since the snapshot was taken includes reading, from the snapshot bitmap, contents of a particular bitmap entry corresponding to the particular data block identified by the write instruction, the new data (i) not being a first update of the particular portion of the file system since the snapshot was taken when the contents of the particular bitmap entry define a first value, and (ii) being a first update of the particular portion of the file system since the snapshot was taken when the contents of the particular bitmap entry define a second value which is different than the first value.

15. An electronic apparatus as in claim 14 wherein the metadata further includes a snapshot blockmap having a set of blockmap entries, each blockmap entry corresponding to a respective data block of the file system;

and wherein the control circuit is further constructed and arranged to;

in response to copying the original data from the cache memory to the snapshot storage, store a pointer in a particular blockmap entry corresponding to the particular data block identified by the write instruction, the pointer being an address of the snapshot storage which stores the original data copied from the cache memory.