Avoiding long access latencies in redundant storage systems
Avoiding long access latencies in redundant storage systems is disclosed, including: determining a first device associated with a request is in a slow access period based at least in part on a scheduled slow access period, wherein the first device is included in a plurality of devices associated with a redundant storage system; and reconstructing data associated with the request from one or more devices included in the plurality of devices other than the first device.
1. A system, comprising:
a processor configured to:
determine that a first device associated with a request is in a slow access period based at least in part on a scheduled slow access period, wherein the first device is included in a plurality of devices associated with a redundant storage system; and
reconstruct data associated with the request from one or more devices included in the plurality of devices other than the first device; and
a storage device configured to store the request.
2. The system of claim 1 , wherein the request includes an identifier associated with the first device.
3. The system of claim 1 , wherein the first device is forced into the slow access period by an issued command associated with triggering a device to enter a slow access period.
4. The system of claim 1 , wherein the processor is further configured to determine a schedule of forced slow access periods corresponding to the plurality of devices, wherein the schedule of forced slow access periods indicates when each device is to be forced into a respective slow access period.
5. The system of claim 4 , wherein the schedule of forced slow access periods does not schedule for two or more devices to be forced into slow access periods at overlapping periods.
6. The system of claim 4 , wherein the schedule of forced slow access periods is configured such that a device does not enter a corresponding scheduled slow access period substantially more often than it would have if it had been allowed to enter the slow access period on its own.
7. The system of claim 1 , wherein the first device and the one or more other devices comprise at least a portion of a redundant array of inexpensive disks (RAID) group.
8. The system of claim 1 , wherein to reconstruct the data associated with the request from one or more devices included in the plurality of devices other than the first device comprises using at least parity data stored on the one or more devices.
9. The system of claim 1 , wherein to determine that the first device associated with the request is in the slow access period is based at least in part on determining that a state associated with the first device indicates that the first device is currently in the slow access period.
10. The system of claim 9 , wherein the processor is further configured to:
determine that the first device is no longer in the slow access period; and
update the state of the first device to indicate that the first device is no longer in the slow access period.
11. A method, comprising:
determining that a first device associated with a request is in a slow access period based at least in part on a scheduled slow access period, wherein the first device is included in a plurality of devices associated with a redundant storage system; and
reconstructing data associated with the request from one or more devices included in the plurality of devices other than the first device.
12. The method of claim 11 , wherein the request includes an identifier associated with the first device.
13. The method of claim 11 , wherein the first device is forced into the slow access period by an issued command associated with triggering a device to enter a slow access period.
14. The method of claim 11 , further comprising determining a schedule of forced slow access periods corresponding to the plurality of devices, wherein the schedule of forced slow access periods indicates when each device is to be forced into a respective slow access period.
15. The method of claim 14 , wherein the schedule of forced slow access periods does not schedule for two or more devices to be forced into slow access periods at overlapping periods.
16. The method of claim 14 , wherein the schedule of forced slow access periods is configured such that a device does not enter a corresponding scheduled slow access period substantially more often than it would have if it had been allowed to enter the slow access period on its own.
17. The method of claim 11 , wherein the first device and the one or more other devices comprise at least a portion of a redundant array of inexpensive disks (RAID) group.
18. The method of claim 11 , wherein reconstructing data associated with the request from the one or more other devices included in the plurality of devices other than the first device comprises using at least parity data stored on the one or more devices.
19. The method of claim 11 , wherein determining that the first device associated with the request is in the slow access period is based at least in part on determining that a state associated with the first device indicates that the first device is currently in the slow access period.
20. A computer program product, the computer program product being embodied in a non-transitory computer readable storage medium and comprising computer instructions for:
determining that a first device associated with a request is in a slow access period based at least in part on a scheduled slow access period, wherein the first device is included in a plurality of devices associated with a redundant storage system; and
reconstructing data associated with the request from one or more devices included in the plurality of devices other than the first device.