Maintaining data access during failure of a controller

A method for maintaining data access during failure of a first controller in a multiple controller storage subsystem, the multiple controller storage subsystem having an array of data storage devices and at least one other controller for managing the data storage, the method comprising:

• in a non-testing mode, the first controller detecting an error in the first controller and thereby initiating a process to maintain data access during failure of the first controller, the process to maintain data access during failure of the first controller comprising:
• the first controller saving its internal state information;
• pausing operation of the at least one other controller;
• and
• the at least one other controller saving its internal state information at the time of pausing without resetting;
• and,
• continuing operation of the at least one other controller, wherein only the first controller resets during the process to maintain data access during failure of the first controller, wherein the internal state information of the first and the at least one other controller is saved to permit diagnosis of the failure of the first controller, wherein one of the first and at least one other controller records the destination of the saved internal state information of an other of the first and at least one other controller, wherein a flag is set when internal status data save operation is occurring to prevent another internal status data save operation from being invoked, wherein the flag is set to prevent the another internal status data save operation from being invoked before the greater time period of a set timeout period and the time period to write the internal status data to a memory, wherein the storage subsystem comprises a Fibre Channel Arbitrated Loop system and the first controller and the at least one other controller comprise host bus adapters.

A storage subsystem comprising at least two controllers for managing data storage, the at least two controllers coupled to at least one data storage device, the storage subsystem further comprising:

• a first controller of the at least two controllers adapted for saving its internal state information during a failure of the first controller in a non-testing mode, the first controller detecting an error in the first controller and thereby initiating a process to maintain data access during failure of the first controller;
• and,
• at least one other controller of the at least two controllers adapted for pausing its operation, and continuing its operation during the failure of the first controller, wherein only the first controller resets during the process to maintain data access during failure of the first controller, wherein the internal state information of the first and the at least one other controller is saved in order to permit diagnosis of the failure of the first controller, wherein a flag is set when internal status data save operation is occurring to prevent another internal status data save operation from being invoked, wherein the flag is set to prevent the another internal status data save operation from being invoked before the greater time period of a set timeout period and the time period to write the internal status data to a memory, wherein the storage subsystem comprises a Fibre Channel Arbitrated Loop system and the at least two controllers comprise host bus adapters.