System and method for fault tolerant controller for network raid

Abstract

A fault-tolerant and efficient way of deducing a set of inconsistent stripes for a network RAID protocol, wherein clients forward input/output (I/O) to a particular controller device called the coordinator, which executes RAID logic and which sends out device IOs to the relevant storage devices. If the coordinator fails then a new coordinator reconstructs its state from the storage devices.

1 . A data storage system defining plural data units, comprising: 
 plural storage controllers in at least one data unit a coordinator being instantiated in one of the controllers, the coordinator implementing logic comprising:    receiving at least one write command representing data from a client;    appending, to the data, data location tags to render at least one tagged device write;    sending the tagged device write to at least one storage device for storage of at least some tags, wherein if the coordinator fails, the tags stored in the storage device are used by a substitute coordinator instantiated on a storage controller in the data unit to repair at least one data inconsistency.    
     
     
         2 . The system of  claim 1 , wherein the location tags represent data storage location information.  
     
     
         3 . The system of  claim 1 , wherein the location tags represent stripe location.  
     
     
         4 . The system of  claim 1  wherein the location tags represent span location.  
     
     
         5 . The system of  claim 1 , wherein the system is implemented in a RAID system.  
     
     
         6 - 9 . (canceled)  
     
     
         10 . A method for promoting fault tolerance in a storage network, comprising: 
 receiving host writes at a coordinator;    transforming each host write into plural device writes, each device write being sent to a respective storage device, each device write including at least one tag;    in the event of coordinator failure, instantiating a substitute coordinator;    using the tags to repair data inconsistencies, if any, between controllers by means of the substitute coordinator; and    periodically discarding tags from the controllers.    
     
     
         11 . The method of  claim 10 , wherein the coordinator serializes host writes from multiple clients.  
     
     
         12 . The method of  claim 10 , wherein the tags represent data storage location information from the client.  
     
     
         13 . The method of  claim 10 , wherein the location tags represent stripe location.  
     
     
         14 . The method of  claim 10 , wherein the location tags represent span location.  
     
     
         15 . The method of  claim 10 , wherein the method is implemented in a RAID system.  
     
     
         16 - 18 . (canceled)  
     
     
         19 . A computer program storage device bearing logic executable by a digital processing apparatus, comprising: 
 means for tagging, with location tags, device writes derived from a host write;    means for sending the device writes to respective storage devices;    means for resolving data inconsistencies between at least two storage devices using the tags; and    means for discarding tags from the storage devices.    
     
     
         20 . The computer program storage device  claim 19 , comprising means for serializing host writes from multiple clients.  
     
     
         21 . The computer program storage device  claim 19 , wherein the tags represent data storage location information from the client.  
     
     
         22 . The computer program storage device  claim 19 , wherein the location tags represent stripe location.  
     
     
         23 . The computer program storage device  claim 19 , wherein the location tags represent span location.  
     
     
         24 . The computer program storage device  claim 19 , wherein the means are implemented in a RAID system.  
     
     
         25 - 27 . (canceled)