Data reliability and availability increase with Panasas® ActiveStor® solution. In NAS systems, adding storage capacity boosts the frequency of drive failures, simply because there are more drives.
Panasas uses software-based erasure codes to protect each file individually instead of protecting whole raw storage devices. File-level protection accelerates rebuild times, increasing data availability and reliability with scale.
The Panasas PanFS® parallel file system advanced per-file, distributed-erasure-coding software implementation offers two levels of protection. It applies dual-parity erasure-coded protection as it distributes data in each file across storage nodes in the cluster. This protection corrects up to two simultaneous failures, whether drives or whole nodes.
The second level of protection handles failures that occur in a portion of a drive, where just a range of blocks within a drive is inaccessible. This allows the ActiveStor appliance to detect and correct sector-level errors and other low-level issues without requiring a full node-level rebuild.
DATA RELIABILITY | VOLUME OF REBUILD DATA | TIME TO REBUILD DATA | EXTENDED FILE SYSTEM AVAILABILITY (EF SA) | |
Legacy NAS w/ RAID 6 | Decreases with scale |
|
|
|
PanFS w/ RAID 6+ | Increases with scale |
|
|
|
The Panasas PanFS® operating environment performs background scrubbing of all files. It checks to ensure that all levels of data-protection information for each file, such as code parity information, match the data in the file.
The vertical parity layer can detect and correct latent “bit rot” errors in the drive media. Data scrubbing helps catch any latent failures early so you can fix them before data loss occurs.
The PanFS storage operating system provides extended file system availability protection for the namespace, directory hierarchy, and file names. In the extremely unlikely event of an unrecoverable failure, the PanFS platform pinpoints affected files and isolates them.
With this additional layer of protection, the PanFS solution can usually identify the full path names of affected files. Users enjoy uninterrupted access to all files unaffected by the failure, and the PanFS system logs any affected files for restoration from backup or another source.
Ensuring data reliability while scale and performance increase requires parallel reconstruction.
Because PanFS rebuilds files instead of drives, and stripes files across differing subsets of all the storage nodes that it manages, it rebuilds all files affected by a failure in parallel.
The PanFS rebuild is not limited by the performance of a single drive. Rebuild performance scales as linearly as data access performance.