Homogeneous recovery in a redundant memory system

A computer program product for performing recovery, comprising a non-transitory tangible storage medium readable by a processing circuit and storing instructions for execution by the processing circuit for performing a method comprising:

• receiving a notification that a memory channel has failed, the memory channel one of a plurality of memory channels in a memory system;
• blocking off new operations from starting on the memory channels;
• shutting down the failing memory channel and ignoring any pending stores and fetches in the failing memory channel;
• based on shutting down the failing memory channel, completing any in-progress fetches only on a portion of the memory channels based on the failing memory channel being detected, the portion of the memory channels consisting of non-failing channels;
• performing a recovery operation on the memory channels, the recovery operation comprising resetting the memory channel and performing data calibration on at least a first subset of the memory channels in the memory system while any other memory channels in the memory system that are not in the first subset are idle;
• retrying any pending stores that were issued prior to receiving notification of the failing memory channel after performing the recovery operation;
• and
• starting the new operations on at least a second subset of the memory channels, the memory system capable of operating with the second subset of the memory channels.

A memory system comprising:

• a memory controller;
• a plurality of memory channels in communication with the memory controller;
• an error detection code mechanism configured for detecting a failing memory channel;
• at least one of a forward progress monitor and an interface monitor for determining actions performed by a recovery operation;
• and
• an error recovery mechanism configured for:
• receiving notification of the failing memory channel;
• blocking off new operations from starting on the memory channels;
• shutting down the failing memory channel and ignoring any pending stores and fetches in the failing memory channel;
• based on shutting down the failing memory channel, completing any in-progress fetches only on a portion of the memory channels based on the failing memory channel being detected, the portion of the memory channels consisting of non-failing channels;
• performing the recovery operation on the memory channels;
• retrying any pending stores that were issued prior to receiving notification of the failing memory channel after performing the recovery operation;
• and
• starting the new operations on at least a first subset of the memory channels, the memory system capable of operating with the first subset of the memory channels.