Granular entry self-healing

A client device of a distributed file system (DFS), the client device comprising:

• a memory;
• and
• a processing device operatively coupled to the memory, the processing device to:
• receive a file operation request;
• request creation of a plurality of entries, wherein each entry of the plurality of entries is created in a corresponding file index in each of a set of servers in the DFS, the plurality of entries corresponding to replicas of directories and one or more files that are a subject of the file operation, each entry of the plurality of entries in the file index corresponding to a respective replica of the replicas of directories and having a unique identifier associated with the respective replica;
• request performance of the file operation by each server in the set of servers of the DFS;
• receive one or more confirmations from the set of servers, each received confirmation corresponding to an indication that the file operation was completed by a corresponding server in the set of servers;
• and
• responsive to each of the one or more confirmations, send a request to a respective server to clear a respective entry from the file index corresponding to a respective confirmation;
• wherein, responsive to a reboot of a first server in the set of servers, a processing device of the first server is to automatically process the file index on that server by performing a corresponding file operation for each entry in the file index, wherein performing the corresponding file operation includes requesting a copy of a file indicated by each entry from a replica in the set of servers.

A non-transitory machine-readable storage medium storing instructions which, when executed, cause a processing device to perform operations comprising:

• receiving, by a client of the DFS, a file operation request;
• requesting, by the client, creation of a plurality of entries, wherein each entry of the plurality of entries is created in a corresponding file index in each of a set of servers in the DFS, the plurality of entries corresponding to replicas of directories and one or more files that are a subject of the file operation, each entry of the plurality of entries in the file index corresponding to a respective replica of the replicas of directories and having a unique identifier associated with the respective replica;
• requesting, by the client, performance of the file operation by each server in the set of servers of the DFS;
• receiving, by the client, one or more confirmations from the set of servers, each received confirmation corresponding to an indication that the file operation was completed by a corresponding server in the set of servers;
• and
• responsive to each of the one or more confirmations, sending a request, by the client, to a respective server to clear a respective entry from the file index corresponding to a respective confirmation;
• wherein, responsive to a reboot of a first server in the set of servers, a processing device of the first server is to automatically process the file index on that server by performing a corresponding file operation for each entry in the file index, wherein performing the corresponding file operation includes requesting a copy of a file indicated by each entry from a replica in the set of servers.