Method and apparatus for predictive failure handling of interleaved dual in-line memory modules
An information handling system includes interleaved dual in-line memory modules (DIMMs) that are partitioned into logical partitions, wherein each logical partition is associated with a namespace. A DIMM controller sets a custom DIMM-level namespace-based threshold to detect a DIMM error and to identify one of the logical partitions of the DIMM error using the namespace associated with the logical partition. The detected DIMM error is repaired if it exceeds an error correcting code (ECC) threshold.
1. An information handling system, comprising:
a plurality of interleaved dual in-line memory modules (DIMMs) that are partitioned into one or more logical partitions, wherein each logical partition is associated with a namespace; and
a DIMM controller coupled to the DIMMs, wherein the DIMM controller is configured to detect a DIMM error and to identify one of the logical partitions of the DIMM error using the namespace associated with the logical partition, and to perform a repair when a number of occurrences of the DIMM error exceeds an error correcting code (ECC) threshold.
2. The information handling system of claim 1 , wherein the namespace includes a physical address and a range of portions allocated to each logical partition.
3. The information handling system of claim 1 , wherein each namespace corresponds to a particular logical partition that is configured on each of the DIMMs by the DIMM controller.
4. The information handling system of claim 1 , wherein a custom DIMM-level namespace-based threshold includes a comparison of the detected DIMM error to the ECC threshold.
5. The information handling system of claim 4 , wherein the DIMM controller performs the repair when the detected DIMM error is within a critical region range of the ECC threshold.
6. The information handling system of claim 1 , wherein the repair includes a mapping of data from an affected channel to another channel within a same or a different DIMM of the same namespace, or remapping the data from the affected channel to another channel within the same or different DIMM of a different namespace.
7. The information handling system of claim 1 , wherein one of the DIMMs include a non-volatile DIMM.
8. The information handling system of claim 7 , wherein the repair is performed during boot time or run time.
9. The information handling system of claim 1 , wherein the DIMM controller is configured to duplicate, modify, transfer, or cancel data within the same associated namespace.
10. A method, comprising:
partitioning a plurality of dual in-line memory modules (DIMMs) into one or more logical partitions, wherein each logical partition is associated with a namespace;
detecting, by a DIMM controller, a DIMM error and identifying one of the logical partitions of the DIMM error using the namespace associated with the logical partition; and
performing a repair in response to a determined impact of the detected DIMM error when the detected DIMM error exceeds an error correcting code (ECC) threshold.
11. The method of claim 10 , wherein the partitioning includes assigning of each logical partition to a different one of the DIMMs.
12. The method of claim 11 , wherein the namespace includes a physical address and a range of portions allocated to each logical partition.
13. The method of claim 10 , further comprising: setting of a custom DIMM-level namespace-based threshold by comparing the detected DIMM error to the ECC threshold.
14. The method of claim 13 , wherein the repair is performed when the detected DIMM error is within a critical region range of the ECC threshold.
15. The method of claim 10 , wherein the repair includes a mapping of data from the associated namespace to another namespace.
16. The method of claim 10 , wherein the plurality of DIMMs include interleaved non-volatile DIMMs.
17. A server comprising:
a plurality of interleaved dual in-line memory modules (DIMMs) that are partitioned into one or more logical partitions, wherein each logical partition is associated with a namespace; and
a DIMM controller coupled to the interleaved DIMMs, wherein the DIMM controller is configured to detect a DIMM error and to identify one of the logical partitions of the DIMM error using the namespace associated with the logical partition, and to perform a repair when the detected DIMM error exceeds an error correcting code (ECC) threshold;
wherein the namespace includes a physical address and a range of portions allocated to each logical partition; and
wherein each namespace corresponds to a particular logical partition.
18. The server of claim 17 , wherein the partitions are configured on the DIMMs by the DIMM controller.
19. The server of claim 18 , wherein the DIMM controller is configured to dynamically modify, replace, reassign, or bypass the configured logical partition on the DIMMs.
20. The server of claim 17 , wherein the interleaved DIMMs include one or more non-volatile DIMMs.