System for speculative block IO aggregation to reduce uneven wearing of SCMs in virtualized compute node by offloading intensive block IOs
An information handling system may include at least one processor and a non-transitory memory coupled to the at least one processor. The information handling system may be configured to implement a disk manager configured to receive a command to perform a block input/output (I/O) job at a first physical storage resource. The disk manager may, based on characteristics of the block I/O job, classify the block I/O job as an intensive write. In response to the block I/O job being classified as an intensive write, the disk manager may invoke a speculative I/O aggregator module of the disk manager. The disk manager may create a delta disk image for the first physical storage resource at a second physical storage resource, and redirect the block I/O job to the delta disk image at the second physical storage resource.
1. An information handling system comprising:
a processor;
a memory communicatively coupled to the processor and comprising a plurality of physical storage resources; and
a disk manager comprising a program of instructions, the disk manager configured to, when read and executed by the processor:
receive a command to perform a block input/output (I/O) job at a first physical storage resource;
based on characteristics of the block I/O job, classify the block I/O job as an intensive write, wherein the classification is based at least in part on a number of bytes or blocks to be written in the block I/O job being above a selected threshold, and wherein the selected threshold is based at least in part on a wear level of the first physical storage resource;
in response to the block I/O job being classified as an intensive write, invoke a speculative I/O aggregator module of the disk manager;
create a delta disk image for the first physical storage resource at a second physical storage resource; and
redirect the block I/O job to the delta disk image at the second physical storage resource such that the block I/O job is applied to the second physical storage resource instead of the first physical storage resource.
2. The information handling system of claim 1 , wherein the command to perform the block I/O job is received from a hypervisor executing on the information handling system.
3. The information handling system of claim 1 , wherein the command to perform the block I/O job is received from a virtualized operating system executing on the information handling system.
4. The information handling system of claim 1 , wherein the first and second physical storage resources comprise storage-class memory (SCM) dual in-line memory modules (DIMMs).
5. The information handling system of claim 1 , further configured to determine at every power-on cycle of the information handling system, information regarding wear leveling, available space, and total capacity for the first and second physical storage resources.
6. The information handling system of claim 1 , wherein the second physical storage resource is selected out of a plurality of potential second physical storage resources based on having a lowest wear level among the plurality of potential second physical storage resources.
7. The information handling system of claim 1 , wherein the disk manager is further configured to:
in response to a user command, merge the delta disk image with the first physical storage resource by committing changes from the delta disk image to the first physical storage resource.
8. The information handling system of claim 1 , wherein the delta disk image is configured to store the redirected block I/O job as XORed data relative to data stored on the first physical storage resource.
9. A method comprising:
receiving, at a disk manager of an information handling system, a command to perform a block input/output (I/O) job at a first physical storage resource of the information handling system;
based on characteristics of the block I/O job, the disk manager classifying the block I/O job as an intensive write, wherein the classification is based at least in part on a number of bytes or blocks to be written in the block I/O job being above a selected threshold, and wherein the selected threshold is based at least in part on a wear level of the first physical storage resource;
in response to the block I/O job being classified as an intensive write, the disk manager invoking a speculative I/O aggregator module;
the disk manager creating a delta disk image for the first physical storage resource at a second physical storage resource; and
the disk manager redirecting the block I/O job to the delta disk image at the second physical storage resource such that the block I/O job is applied to the second physical storage resource instead of the first physical storage resource.
10. The method of claim 9 , further comprising:
redirected at least one second block I/O job to the delta disk image at the second physical storage resource.
11. An article of manufacture comprising a non-transitory, computer-readable medium having computer-executable code thereon that is executable by a processor of an information handling system for implementing a disk manager configured to:
receive a command to perform a block input/output (I/O) job at a first physical storage resource;
based on characteristics of the block I/O job, classify the block I/O job as an intensive write, wherein the classification is based at least in part on a number of bytes or blocks to be written in the block I/O job being above a selected threshold, and wherein the selected threshold is based at least in part on a wear level of the first physical storage resource;
in response to the block I/O job being classified as an intensive write, invoke a speculative I/O aggregator module of the disk manager;
create a delta disk image for the first physical storage resource at a second physical storage resource; and
redirect the block I/O job to the delta disk image at the second physical storage resource such that the block I/O job is applied to the second physical storage resource instead of the first physical storage resource.
12. The article of claim 11 , wherein the command to perform the block I/O job is received from a hypervisor executing on the information handling system.
13. The article of claim 11 , wherein the command to perform the block I/O job is received from a virtualized operating system executing on the information handling system.
14. The article of claim 11 , wherein the first and second physical storage resources comprise storage-class memory (SCM) dual in-line memory modules (DIMMs).
15. The article of claim 11 , wherein the processor is further configured to determine at every power-on cycle of the information handling system, information regarding wear leveling, available space, and total capacity for the first and second physical storage resources.
16. The article of claim 11 , wherein the second physical storage resource is selected out of a plurality of potential second physical storage resources based on having a lowest wear level among the plurality of potential second physical storage resources.
17. The article of claim 11 , wherein the disk manager is further configured to:
in response to a user command, merge the delta disk image with the first physical storage resource by committing changes from the delta disk image to the first physical storage resource.
18. The article of claim 17 , wherein the reconciling is performed in response to an indication that a wear level of the second physical storage resource has reached a threshold level.