Data reconstruction in distributed storage systems

A method of operating a distributed storage system, the method comprising:

• identifying, by data processing hardware, unhealthy chunks within a given stripe of a file, the file divided into stripes, each stripe comprising data chunks and non-data chunks formed by a nested coding process, the data chunks and non-data chunks arranged in an array of rows and columns, the non-data chunks comprising different types of check chunks that include at least one code-check chunk associated with each column in the array of rows and columns, word-check chunks associated with a row in the array of rows and columns, and code-check-word-check chunks associated with other check chunks, wherein an error-correcting functionality of each check chunk among the different types of check chunks of the non-data chunks is defined based on a location of the corresponding check chunk in a data structure of a corresponding stripe;
• for each unhealthy chunk, determining, by the data processing hardware, whether the unhealthy chunk is an unhealthy data chunk or an unhealthy non-data chunk;
• identifying, by the data processing hardware, healthy chunks comprising healthy data chunks and healthy non-data chunks available for reconstructing the unhealthy chunks;
• and
• reconstructing, by the data processing hardware, unhealthy data chunks before reconstructing unhealthy non-data chunks using the available healthy chunks,
• wherein when the unhealthy chunk is an unhealthy word-check chunk:
• reconstructing the unhealthy word-check chunk using healthy word-check chunks and healthy code-check-word-check chunks;
• determining, by the data processing hardware, whether reconstruction of the unhealthy word-check chunk is possible using only healthy word-check chunks and healthy code-check-word-check chunks;
• and
• when reconstruction of the unhealthy word-check chunk is not possible using only healthy word-check chunks and healthy code-check-word-check chunks, reconstructing, by the data processing hardware, the unhealthy word-check chunk using any healthy chunks.