Emc IP Holding Company LLC

Pengju Shang

George Mathew

Pranay Singh

Murthy V Mamidi

Englin Koay

Dhawal Bhagwat

In one example, a method for writing data includes receiving a write request and performing a first type of logging process in connection with the write request, and creating a corresponding first logging record. Additionally, a second type of logging process is performed in connection with the write request, and a corresponding second logging record created, where the second type of logging process is different from the first type of logging process. Next, a determination is made, as between the two logging records, which of the logging records requires the least amount of non-volatile random access memory (NVRAM), and the logging record that requires the least amount of NVRAM is written to the NVRAM.

Hybrid NVRAM logging in filesystem namespace

In one example, a method includes receiving metadata in the form of a modification to metadata represented by a file system namespace abstraction, wherein the file system namespace abstraction corresponds to less than an entire file system namespace, and the file system namespace abstraction includes one or more pages, and one of the pages corresponds to a particular cached block, updating the file system namespace abstraction based on the received metadata, determining if caching is enabled for the file system namespace abstraction, when caching is enabled for the file system namespace abstraction, caching the updated file system namespace abstraction in SSD storage that includes the cached block, and maintaining a status of the cached block in a Dtable of the SSD storage.

Namespace performance acceleration by selective SSD caching

In one example, a method includes receiving metadata in the form of a modification to metadata represented by a file system namespace abstraction, and the file system namespace abstraction corresponds to less than the entire file system namespace. Next, the file system namespace abstraction is updated based on the received metadata. Next, a determination is made whether or not caching is enabled for the file system namespace abstraction. If caching is enabled for the file system namespace abstraction, the updated file system namespace abstraction is cached in SSD storage.

According to some embodiment, a storage system selects one or more directories within a file system as candidates for caching based on directory statistics associated with the directories, where each of the directories includes one or more file objects stored in the storage system. For each of the selected directories, the system determines whether the directory is to be cached based on a directory cache policy. The system caches the directory in a cache memory device in response to determining that the directory is to be cached.

Method and system for caching directories in a storage system

Embodiments are described for a method and system for improving B+Tree scan performance by receiving a data access instruction that specifies pages to be accessed in a data store utilizing a B+Tree data structure; defining a read-ahead context comprising an array of page numbers corresponding to the specified pages; loading the read-ahead context array into a read-ahead cache; and reading the first page of the read-ahead context in a synchronous manner, and each of the subsequent pages of the read-ahead context in an asynchronous manner.

Context-based read-ahead for B+ tree data structures in a deduplication system

In one embodiment, a computer-implemented method includes capturing a consistent state of data blocks in a namespace cache of a deduplicating storage system. The data blocks contains data for a file system namespace organized in a hierarchical data structure. Each leaf page of the hierarchical data structure contains one or more data blocks. The method further includes determining, for each data block, whether the data block has been written to base on the captured consistent state. For at least one of the written data blocks in the namespace cache, the method includes searching, in the hierarchical data structure, adjacent data blocks to find in the namespace cache one or more data blocks that have also been written to, and upon finding the one or more adjacent written data blocks, flushing the written data block and the found one or more adjacent written data blocks together into a common storage unit.

Method and system for locality based cache flushing for file system namespace in a deduplicating storage system

In some embodiments, in response to receiving an indication to warm up a SSD cache, a system identifies namespace data of a file system to be warmed up separated from content data. The system identifies one or more namespace pages of the namespace data. For each of the one or more namespace pages, the system locks the namespace page in a read-only mode and determines if the namespace page is dirty. If the namespace page is dirty, the system releases the namespace page from the read-only mode without copying the namespace page to the SSD cache. If the page is clean (i.e., not dirty) then the system copies the namespace page to the SSD cache and releases the namespace page from the read-only mode to reduce cache misses of the namespace on the SSD cache.

Smart namespace SSD cache warmup for storage systems

A computer-implemented method is implemented to compress a file system namespace. The method starts with determining that a first leaf page of a hierarchical data structure of a storage system is sparse. The hierarchical data structure contains root node layer, intermediate node layer, and leaf page layer, where each leaf page contains namespace entries corresponding to a file or a directory of a file system, and where being sparse means a first size of the entries of the first leaf page is less than a first percentage. The method continues with identifying a second immediate adjacent leaf page as a merging candidate. The method continues with merging the second leaf page into the first leaf page upon determining that a second size of combined entries of the first and second leaf pages is less than a second percentage larger than the first percentage.

Growing community of inventors

Milpitas, CA, United States of America

Pengju Shang