Lock state reconstruction for non-disruptive persistent operation

A first storage node for use within a high-availability network storage cluster, the first storage node comprising:

• a processor;
• a network interface through which the storage node can communicate with other nodes within the high-availability network storage cluster;
• a data storage device;
• and
• a memory storing instructions which, when executed by the processor, cause the storage node to perform a process including:
• reconstructing, at the first storage node after an event, a lock data structure representing a lock for a data container, based on a mirrored lock data structure stored at a second storage node of a storage cluster;
• wherein the mirrored lock data structure includes metadata of an original lock data structure associated with the lock for providing lock state information for the reconstructing;
• setting an indicator by the first storage node associated with the lock data structure to indicate that the lock data structure has been reconstructed;
• wherein the indicator is a flag;
• determining by the first storage node that the lock data structure has been recently reconstructed based on whether the flag associated with the lock data structure exists, in response to receiving an operation request from a client device to perform an operation on the data container;
• terminating, by the first storage node, a connection between the first storage node and the client device, in response to determining that the lock data structure representing the lock has been recently reconstructed by the first storage node, and the flag associated with the lock data structure is set;
• removing the lock data structure by the first storage node when the storage node receives no reconnect request from the client device, in response to terminating the storage connection, after a predetermined time period;
• and
• clearing the flag associated with the lock data structure in response to a reconnect request from the client device.

A non-transitory machine-readable medium having stored thereon instructions for performing a method comprising machine executable code which when executed by at least one machine, causes the machine to:

• reconstruct, at a first storage node after an event, a lock data structure representing a lock for a data container, based on a mirrored lock data structure stored at a second storage node of a storage cluster;
• wherein the mirrored lock data structure includes metadata of an original lock data structure associated with the lock for providing lock state information for the reconstructing;
• set an indicator by the first storage node associated with the lock data structure to indicate that the lock data structure has been reconstructed;
• wherein the indicator is a flag
• determine by the first storage node that the lock data structure has been recently reconstructed based on whether the flag associated with the lock data structure exists, in response to receiving an operation request from a client device to perform an operation on the data container;
• terminate, by the first storage node, a connection between the first storage node and the client device, in response to determining that the lock data structure representing the lock has been recently reconstructed by the first storage node, and the flag associated with the lock data structure is set;
• remove the lock data structure by the first storage node when the storage node receives no reconnect request from the client device, in response to terminating the storage connection, after a predetermined time period;
• and
• clear the flag associated with the lock data structure in response to a reconnect request from the client device.