Resource allocation in a storage system

Abstract

A method for resource allocation in a storage system, the method may include determining to reduce a load imposed on a compute core of the storage system, wherein the compute core is responsible for executing (a) one or more background task shards, and (b) at least one non-sharded task for responding to an I/O request sent to the storage system; and reducing the load imposed on the compute core by reallocating at least one background task shard of the one or more background task shards to another compute core of the storage system.

1 . A method for resource allocation in a storage system, the method comprising:
 determining to reduce a load imposed on a compute core of the storage system, wherein the compute core is responsible for executing (a) one or more background task shards, and (b) at least one non-sharded task for responding to an I/O request sent to the storage system; and   reducing the load imposed on the compute core by reallocating at least one background task shard of the one or more background task shards, being executed by the compute core, to another compute core of the storage system that is responsible for executing at least one other non-sharded task for responding to another I/O request sent to the storage system;   wherein the determining includes limiting a number of background task shards reallocated from the compute node within a certain time window.   
     
     
         2 . The method according to  claim 1  comprising preventing from reallocating one of the at least one non-sharded task to another compute core. 
     
     
         3 . (canceled) 
     
     
         4 . The method according to  claim 1  wherein the reallocating comprises selecting at random the another compute core. 
     
     
         5 . The method according to  claim 1  wherein the determining to reduce the load imposed on the compute core follows an addition of one or more added compute cores to the storage system. 
     
     
         6 . The method according to  claim 5  comprising reallocating the at least one background task shard to an added compute core of the storage system. 
     
     
         7 . The method according to  claim 1  wherein the determining to reduce the load imposed on the compute core is a part of a load balancing process for balancing loads imposed on multiple processing nodes of the storage system. 
     
     
         8 . The method according to  claim 1 , wherein the one or more background task shards are related to tasks that differ from responding to I/O requests sent to the storage system. 
     
     
         9 . The method according to  claim 1  wherein the reallocating comprises selecting the at least one background task shard based on an age of the at least one background task shard and based on an age of a non-selected background task shard of the one or more background task shards, wherein the age of the at least one background task shard represents a time lapsed from a reception of the at least one background task shard by the compute core, and the age of the non-selected background task shard represents a time lapsed from a reception of the non-selected background task shard by the compute core. 
     
     
         10 . The method according to  claim 1  wherein the reallocating comprises prioritizing reallocating older background task shards over reallocating of younger background task shards, wherein the younger background task shards are being executed by the compute core during a shorter time than the older background task shards. 
     
     
         11 . The method according to  claim 1  wherein the another compute core has a load that is lower than the load of the compute core. 
     
     
         12 . The method according to  claim 1  wherein the one or more background task shards belong to one or more background tasks that comprise a certain background task that is related to a content written to the storage system, the certain background task is being executed after the writing of the content to the storage system. 
     
     
         13 . The method according to  claim 1  wherein the reallocating is responsive to loads associated with execution of the one or more background task shards. 
     
     
         14 . The method according to  claim 1  wherein a response time of the storage system is more sensitive to (a) an execution of the I/O request, than to (b) an execution of the one or more background task shards. 
     
     
         15 . The method according to  claim 1  comprising assigning at least one out of (a) an additional non-sharded task and (b) an additional background task shard, to the compute core. 
     
     
         16 . The method according to  claim 15  wherein the assigning follows a removal of one or more other compute cores from the storage system. 
     
     
         17 . A storage system comprising multiple compute nodes that comprise compute cores, multiple storage nodes and a load balancer, wherein the load balancer is configured to:
 determine to reduce a load imposed on a compute core of the storage system, wherein the compute core is responsible for executing (a) one or more background task shards, and (b) at least one non-sharded task for responding to an I/O request sent to the storage system; and   reduce the load imposed on the compute core by reallocating at least one background task shard of the one or more background task shards, being executed by the compute core, to another compute core of the storage system that is responsible for executing at least one other non-sharded task for responding to another I/O request sent to the storage system;   wherein the determining includes limiting a number of background task shards reallocated from the compute node within a certain time window.   
     
     
         18 . A non-transitory computer readable medium for resource allocation in a storage system, the non-transitory computer readable medium stores instructions for:
 determine to reduce a load imposed on a compute core of the storage system, wherein the compute core is responsible for executing (a) one or more background task shards, and (b) at least one non-sharded task for responding to an I/O request sent to the storage system; and   reducing the load imposed on the compute core by reallocating at least one background task shard of the one or more background task shards, being executed by the compute core, to another compute core of the storage system that is responsible for executing at least one other non-sharded task for responding to another I/O request sent to the storage system;   wherein the determining includes limiting a number of background task shards reallocated from the compute node within a certain time window.