Fast multipath failover
A host device is configured to obtain a default timeout value of the host device for the submission of an input-output (IO) operation to a storage system and to determine a first timeout value that is less than the default timeout value. The host device is further configured to submit the IO operation to the storage system along a first path using the first timeout value and to determine that the submission of the IO operation along the first path has timed out. The host device is further configured to determine a second timeout value that is greater than the first timeout value and to submit the IO operation to the storage system along a second path using the second timeout value.
1. An apparatus comprising:
a host device comprising a processor coupled to memory, the host device being configured:
to obtain a default timeout value of the host device for submission of an input-output operation to a storage system;
to determine a first timeout value, the first timeout value being less than the default timeout value;
to submit the input-output operation to the storage system along a first path using the first timeout value;
to determine that the submission of the input-output operation along the first path has timed out based at least in part on the first timeout value;
to determine a second timeout value, the second timeout value being greater than the first timeout value;
to submit the input-output operation to the storage system along a second path using the second timeout value based at least in part on the determination that the submission of the input-output operation along the first path has timed out; and
to resubmit the input-output operation to the storage system along the first path using the default timeout value based at least in part on a determination that the submission of the input-output operation along the first path and the second path have timed out;
wherein at least one of determining a first timeout value and determining a second timeout value comprises selecting a number of attempts to submit the input-output operations along the respective first and second paths; and
wherein the number of attempts is greater than one attempt.
2. The apparatus of claim 1 wherein the second timeout value and the default timeout value are the same.
3. The apparatus of claim 1 wherein the second timeout value is less than the default timeout value.
4. The apparatus of claim 3 wherein the host device is further configured:
to determine that the submission of the input-output operation along the second path has timed out based at least in part on the second timeout value;
to determine a third timeout value, the third timeout value being greater than the second timeout value; and
to submit the input-output operation to the storage system along a third path using the third timeout value based at least in part on the determination that the submission of the input-output operation along the second path has timed out.
5. The apparatus of claim 4 wherein the third timeout value and the default timeout value are the same.
6. The apparatus of claim 1 wherein the default timeout value is specified by one of an operating system of the host device and a hardware bus adapter of the host device.
7. The apparatus of claim 1 wherein determining the first timeout value comprises determining the first timeout value based at least in part on one or more of a determined latency associated with the storage system, a prior number of attempts to submit the input-output operation to the storage system, a recent latency on one or more paths along which the input-output operation may be submitted to the storage system, a size of the input-output operation, one or more recent timeouts on the one or more paths along which the input-output operation may be submitted to the storage system and a number of alive paths along which the input-output operation may be submitted to the storage system.
8. The apparatus of claim 1 wherein the host device is further configured:
to determine an indication of at least one of a storage device of the storage system to which the input-output operation is being submitted, a type of the storage device of the storage system to which the input-output operation is being submitted and an application requesting the submission of the input-output operation to the storage system; and
to obtain a preset timeout value corresponding to the determined indication; and
wherein determining first timeout value comprises determining the first timeout value as the obtained preset timeout value.
9. A method comprising:
obtaining a default timeout value of a host device for submission of an input-output operation to a storage system;
determining a first timeout value, the first timeout value being less than the default timeout value;
submitting the input-output operation to the storage system along a first path using the first timeout value;
determining that the submission of the input-output operation along the first path has timed out based at least in part on the first timeout value;
determining a second timeout value, the second timeout value being greater than the first timeout value;
submitting the input-output operation to the storage system along at least a second path using the second timeout value based at least in part on the determination that the submission of the input-output operation along the first path has timed out; and
resubmitting the input-output operation to the storage system along the first path using the default timeout value based at least in part on a determination that the submission of the input-output operation along the first path and the second path have timed out;
wherein at least one of determining a first timeout value and determining a second timeout value comprises selecting a number of attempts to submit the input-output operations along the respective first and second paths;
wherein the number of attempts is greater than one attempt; and
wherein the method is performed by at least one processor of a host device.
10. The method of claim 9 wherein the second timeout value and the default timeout value are the same.
11. The method of claim 9 wherein the second timeout value is less than the default timeout value.
12. The method of claim 11 wherein the method further comprises:
determining that the submission of the input-output operation along the second path has timed out based at least in part on the second timeout value;
determining a third timeout value, the third timeout value being greater than the second timeout value; and
submitting the input-output operation to the storage system using the third timeout value based at least in part on the determination that the submission of the input-output operation along the second path has timed out.
13. The method of claim 12 wherein the third timeout value and the default timeout value are the same.
14. The method of claim 9 wherein determining the first timeout value comprises determining the first timeout value based at least in part on one or more of a determined latency associated with the storage system, a prior number of attempts to submit the input-output operation to the storage system, a recent latency on one or more paths along which the input-output operation may be submitted to the storage system, a size of the input-output operation, one or more recent timeouts on the one or more paths along which the input-output operation may be submitted to the storage system and a number of alive paths along which the input-output operation may be submitted to the storage system.
15. The method of claim 9 wherein the method further comprises:
determining an indication of at least one of a storage device of the storage system to which the input-output operation is being submitted, a type of the storage device of the storage system to which the input-output operation is being submitted and an application requesting the submission of the input-output operation to the storage system; and
obtaining a preset timeout value corresponding to the determined indication; and
wherein determining first timeout value comprises determining the first timeout value as the obtained preset timeout value.
16. A computer program product comprising a non-transitory processor-readable storage medium having stored therein program code of one or more software programs, wherein the program code, when executed by at least one processor of a host device, causes the host device:
to obtain a default timeout value of the host device for submission of an input-output operation to a storage system;
to determine a first timeout value, the first timeout value being less than the default timeout value;
to submit the input-output operation to the storage system along a first path using the first timeout value;
to determine that the submission of the input-output operation along the first path has timed out based at least in part on the first timeout value;
to determine second timeout value, the second timeout value being greater than the first timeout value;
to submit the input-output operation to the storage system along a second path using the second timeout value based at least in part on the determination that the submission of the input-output operation along the first path has timed out; and
to resubmit the input-output operation to the storage system along the first path using the default timeout value based at least in part on a determination that the submission of the input-output operation along the first path and the second path have timed out;
wherein at least one of determining a first timeout value and determining a second timeout value comprises selecting a number of attempts to submit the input-output operations along the respective first and second paths; and
wherein the number of attempts is greater than one attempt.
17. The computer program product of claim 16 wherein the second timeout value and the default timeout value are the same.
18. The computer program product of claim 16
wherein the second timeout value is less than the default timeout value; and
wherein the host device is further configured:
to determine that the submission of the input-output operation along the second path has timed out based at least in part on the second timeout value;
to determine a third timeout value, the third timeout value being greater than the second timeout value; and
to submit the input-output operation to the storage system using the third timeout value based at least in part on the determination that the submission of the input-output operation along the second path has timed out.
19. The computer program product of claim 16 wherein the program code further causes the host device:
to determine an indication of at least one of a storage device of the storage system to which the input-output operation is being submitted, a type of the storage device of the storage system to which the input-output operation is being submitted and an application requesting the submission of the input-output operation to the storage system; and
to obtain a preset timeout value corresponding to the determined indication; and
wherein determining first timeout value comprises determining the first timeout value as the obtained preset timeout value.
20. The apparatus of claim 4 wherein the host device is further configured:
to submit the input-output operation to the storage system along the first path using the default timeout value based at least in part on the determination that the submission of the input-output operation along the third path has timed out.