Map-reduce ready distributed file system
A map-reduce compatible distributed file system that consists of successive component layers that each provide the basis on which the next layer is built provides transactional read-write-update semantics with file chunk replication and huge file-create rates. Containers provide the fundamental basis for data replication, relocation, and transactional updates. A container location database allows containers to be found among all file servers, as well as defining precedence among replicas of containers to organize transactional updates of container contents. Volumes facilitate control of data placement, creation of snapshots and mirrors, and retention of a variety of control and policy information. Also addressed is the use of distributed transactions in a map-reduce system; the use of local and distributed snapshots; replication, including techniques for reconciling the divergence of replicated data after a crash; and mirroring.
1. A non-transitory storage medium comprising instructions that upon execution cause a system to:
maintain, in a container location database (CLDB), information about where each of a plurality of containers is located among a plurality of cluster nodes;
create a snapshot of a file system volume comprising directories and files spread over the plurality of containers, wherein the creating of the snapshot of the file system volume comprises:
creating a snapshot of a name container, wherein the name container comprises metadata for the file system volume, and
after creating the snapshot of the name container, creating snapshots of data containers to produce a plurality of snapshot containers, the data containers storing data of the directories and the files of the file system volume;
create mapping information that maps references to the data containers into references to corresponding snapshot containers of the plurality of snapshot containers;
insert the mapping information into the snapshot of the name container; and
use the mapping information to translate a reference to a given data container to a reference to a corresponding snapshot container.
2. The non-transitory storage medium of claim 1 , wherein the mapping information comprises a table that maps the references to the data containers into the references to the corresponding snapshot containers, and
wherein the inserting comprises inserting, into the snapshot of the name container, the mapping information comprising the table that maps the references to the data containers into the references to the corresponding snapshot containers.
3. The non-transitory storage medium of claim 1 , wherein the instructions upon execution cause the system to:
insert the mapping information into each snapshot container of the plurality of snapshot containers.
4. The non-transitory storage medium of claim 1 , wherein the instructions upon execution cause the system to:
in response to a modification of data blocks in the file system volume, write the modified data blocks to new storage locations, and leave data in the plurality of snapshot containers unchanged.
5. The non-transitory storage medium of claim 1 , wherein the creating of the mapping information occurs after the creating of the snapshot of the name container and the creating of the snapshots of the data containers.
6. The non-transitory storage medium of claim 1 , wherein the CLDB and the plurality of containers including the snapshot of the name container and the plurality of snapshot containers are part of a distributed file system implemented across the plurality of cluster nodes.
7. The non-transitory storage medium of claim 6 , wherein the instructions upon execution cause the system to:
perform a distributed transaction comprising a distributed update on the file system volume, wherein the distributed update is to update data of the file system volume in multiple cluster nodes of the plurality of cluster nodes.
8. A system comprising:
a plurality of cluster nodes;
a processor; and
a non-transitory storage medium storing instructions executable on the processor to:
maintain, in a container location database (CLDB), information about where each of a plurality of containers is located among the plurality of cluster nodes;
create a snapshot of a file system volume comprising directories and files spread over the plurality of containers, wherein the creating of the snapshot of the file system volume comprises:
creating a snapshot of a name container, wherein the name container comprises metadata for the file system volume, and
after creating the snapshot of the name container, creating snapshots of data containers to produce a plurality of snapshot containers, the data containers storing data of the directories and the files of the file system volume;
create mapping information that maps identifiers of the data containers into identifiers of corresponding snapshot containers of the plurality of snapshot containers;
insert the mapping information into the snapshot of the name container; and
use the mapping information to translate an identifier of a given data container to an identifier of a corresponding snapshot container.
9. The system of claim 8 , wherein the mapping information comprises a table that maps the identifiers of the data containers into the identifiers of the corresponding snapshot containers, and
wherein the inserting comprises inserting, into the snapshot of the name container, the mapping information comprising the table that maps the identifiers of the data containers into the identifiers of the corresponding snapshot containers.
10. The system of claim 8 , wherein the metadata in the name container comprises a file identifier map including file identifiers that refer to files containing data of the data containers.
11. The system of claim 8 , wherein the instructions are executable on the processor to:
in response to a modification of data blocks in the file system volume, write the modified data blocks to new storage locations, and leave data in the plurality of snapshot containers unchanged.
12. The system of claim 8 , wherein the creating of the mapping information occurs after the creating of the snapshot of the name container and the creating of the snapshots of the data containers.
13. The system of claim 8 , wherein the CLDB and the plurality of containers are part of a distributed file system implemented across the plurality of cluster nodes.
14. The system of claim 13 , wherein the instructions are executable on the processor to:
perform a distributed transaction comprising a distributed update on the file system volume, wherein the distributed update is to update data of the file system volume in multiple cluster nodes of the plurality of cluster nodes.
15. A method of a system, comprising:
maintaining, in a container location database (CLDB), information about where each of a plurality of containers is located among a plurality of cluster nodes;
creating a snapshot of a file system volume comprising directories and files spread over the plurality of containers, wherein the creating of the snapshot of the file system volume comprises:
creating a snapshot of a name container, wherein the name container comprises metadata for the file system volume, and
after creating the snapshot of the name container, creating snapshots of data containers to produce a plurality of snapshot containers, the data containers storing data of the directories and the files of the file system volume;
creating mapping information that maps references to the data containers into references to corresponding snapshot containers of the plurality of snapshot containers;
inserting the mapping information into the snapshot of the name container; and
using the mapping information to translate a reference to a given data container to a reference to a corresponding snapshot container.
16. The method of claim 15 , further comprising:
inserting the mapping information into each snapshot container of the plurality of snapshot containers.
17. The method of claim 15 , further comprising:
in response to a modification of data blocks in the file system volume, writing the modified data blocks to new storage locations, and leaving data in the plurality of snapshot containers unchanged.
18. The method of claim 15 , wherein the creating of the mapping information occurs after the creating of the snapshot of the name container and the creating of the snapshots of the data containers.
19. The method of claim 15 , wherein the mapping information comprises a table that maps the references to the data containers into the references to the corresponding snapshot containers, and
wherein the inserting comprises inserting, into the snapshot of the name container, the mapping information comprising the table that maps the references to the data containers into the references to the corresponding snapshot containers.
20. The method of claim 15 , wherein the CLDB and the plurality of containers are part of a distributed file system implemented across the plurality of cluster nodes, the method further comprising:
performing a distributed transaction comprising a distributed update on the file system volume, wherein the distributed update is to update data of the file system volume in multiple cluster nodes of the plurality of cluster nodes.