Method and apparatus for migrating volumes and virtual disks

In a distributed storage system having a plurality of storage controllers including first, second and third storage controllers coupled to at least one initiator, the plurality of storage controllers each having a unique IP address, a method for continuing services provided by the first storage controller in response to a failure of the first storage controller, comprising:

• migrating a first virtual disk from the first storage controller to the second storage controller;
• migrating a first volume from the first storage controller to one of the second and third storage controllers;
• and
• updating one of the plurality of storage controllers to redirect communications from the at least one initiator and intended for the first volume to one of the second and third storage controllers, the one of the plurality of storage controllers having a virtual IP address for contacting by the at least one initiator in response to the failure of the first storage controller, the virtual IP address being different from the unique IP address of any of the plurality of storage controllers.

In a distributed storage system having a plurality of storage controllers including first and second storage controllers, a method for transferring at least a portion of a load being handled by the first storage controller, comprising:

• signaling at least some of the plurality of storage controllers to stop sending new requests addressed to a virtual disk on the first storage controller;
• completing outstanding requests for the virtual disk received at the first storage controller;
• migrating the virtual disk from the first storage controller to the second storage controller;
• redirecting each of the plurality of storage controllers using the virtual disk to communicate with the migrated virtual disk on the second storage controller;
• and
• wherein the second storage controller has a protocol stack including from top to bottom a SCBTP target portion, a virtual disk mapping portion, and a DEBTP initiator portion, and the migrating step includes configuring the protocol stack upwardly for the virtual disk.

A method for load balancing in a distributed storage system having a plurality of storage controllers each having a unique IP address, the plurality of storage controllers including first and second storage controllers coupled to at least one initiator, comprising:

• discarding new requests for a volume at the first storage controller while waiting for outstanding requests for the volume to complete;
• migrating the volume from the first storage controller to the second storage controller;
• updating one of the plurality of storage controllers to redirect requests for the volume to the second storage controller, the one of the plurality of storage controllers having a virtual IP address for contacting by the at least one initiator in response to receiving a logout request from the first storage controller, the virtual IP address being different from the unique IP address of any of the plurality of storage controllers;
• and
• requesting the at least one initiator to logout from the volume.

A first storage controller among a plurality of storage controllers coupled to at least one initiator in a distributed storage system, the plurality of storage controllers each having a unique IP address, the first storage controller comprising a processor and a computer readable medium storing therein computer readable instructions that when executed by the processor cause the first storage controller to perform a method for continuing services provided by a second storage controller that has failed, the instructions comprising:

• instructions for migrating a virtual disk from the second storage controller to the first storage controller in response to the first storage controller being selected to handle the virtual disk;
• instructions for migrating a volume from the second storage controller to the first storage controller in response to the first storage controller being selected to handle the volume;
• and
• instructions for updating one of the plurality of storage controllers to redirect communications from the at least one initiator and intended for the volume to the first storage controller in response to the volume being migrated to the first storage controller, the one of the plurality of storage controllers having a virtual IP address for contacting by the at least one initiator in response to failing to connect to the second storage controller, the virtual IP address being different from the unique IP address of any of the plurality of storage controllers.

A first storage controller among a plurality of storage controllers coupled to at least one initiator in a distributed storage system, each of the plurality of storage controllers having a unique IP address and including at least one volume, the first storage controller comprising a processor and a computer readable medium storing therein computer readable instructions that when executed by the processor cause the first storage controller to act as a contact storage controller for the at least one initiator, the instructions comprising:

• instructions to receive a request to access a volume directed to a virtual IP address from an initiator, the virtual IP address being different from the unique IP address of any of the plurality of storage controllers;
• instructions for determining a destination storage controller that has the volume;
• and
• instructions for redirecting the initiator to send the request to the destination storage controller.