Error Recovery in a Storage Cluster

[0001] Solid-state memory, such as flash, is currently in use in solid-state drives (SSD) to augment or replace conventional hard disk drives (HDD), writable CD (compact disk) or writable DVD (digital versatile disk) drives, collectively known as spinning media, and tape drives, for storage of large amounts of data. Flash and other solid-state memories have characteristics that differ from spinning media. Yet, many solid-state drives are designed to conform to hard disk drive standards for compatibility reasons, which makes it difficult to provide enhanced features or take advantage of unique aspects of flash and other solid-state memory. Errors can result from a read failure, or a failure of a component or a subsystem, and a solid-state drive can lose data and/or functionality. Downtime of a system, while a solid-state drive is replaced or repaired, is disruptive, as is data loss.

[0002] It is within this context that the embodiments arise.

[0003] In some embodiments, a plurality of storage nodes within a single chassis is provided. The plurality of storage nodes is configured to communicate together as a storage cluster, each of the plurality of storage nodes having a non-volatile solid-state storage for user data storage, the non-volatile solid state storage including flash memory. The plurality of storage nodes is configured to distribute the user data and metadata associated with the user data throughout the plurality of storage nodes, with erasure coding of the user data. The plurality of storage nodes is configured to recover from failure of two of the plurality of storage nodes by applying the erasure coding to reading the user data from a remainder of the plurality of storage nodes. The plurality of storage nodes is configured to detect an error and engage in an error recovery via one of a processor of one of the plurality of storage nodes, a processor of the non-volatile solid state storage, or the flash memory.

[0004] Other aspects and advantages of the embodiments will become apparent from the following detailed description taken in conjunction with the accompanying drawings which illustrate, by way of example, the principles of the described embodiments.

[0015] The embodiments below describe a storage cluster that stores user data, such as user data originating from one or more user or client systems or other sources external to the storage cluster. The storage cluster distributes user data across storage nodes housed within a chassis, using erasure coding and redundant copies of metadata. Erasure coding refers to a method of data protection in which data is broken into fragments, expanded and encoded with redundant data pieces and stored across a set of different locations, such as disks, storage nodes or geographic locations. Flash memory is one type of solid-state memory that may be integrated with the embodiments, although the embodiments may be extended to other types of solid-state memory or other storage medium, including non-solid state memory. Control of storage locations and workloads are distributed across the storage locations in a clustered peer-to-peer system. Tasks such as mediating communications between the various storage nodes, detecting when a storage node has become unavailable, and balancing I/Os (inputs and outputs) across the various storage nodes, are all handled on a distributed basis. Data is laid out or distributed across multiple storage nodes in data fragments or stripes that support data recovery in some embodiments. Ownership of data can be reassigned within a cluster, independent of input and output patterns. This architecture described in more detail below allows a storage node in the cluster to fail, with the system remaining operational, since the data can be reconstructed from other storage nodes and thus remain available for input and output operations. In various embodiments, a storage node may be referred to as a cluster node, a blade, or a server.

[0016] The storage cluster is contained within a chassis, i.e., an enclosure housing one or more storage nodes. A mechanism to provide power to each storage node, such as a power distribution bus, and a communication mechanism, such as a communication bus that enables communication between the storage nodes are included within the chassis. The storage cluster can run as an independent system in one location according to some embodiments. In one embodiment, a chassis contains at least two instances of the power distribution and the internal and external communication bus which may be enabled or disabled independently. The internal communication bus may be an Ethernet bus, however, other technologies such as Peripheral Component Interconnect (PCI) Express, InfiniBand, and others, are equally suitable. The chassis provides a port for an external communication bus for enabling communication between multiple chassis, directly or through a switch, and with client systems. The external communication may use a technology such as Ethernet, InfiniBand, Fibre Channel, etc. In some embodiments, the external communication bus uses different communication bus technologies for inter-chassis and client communication. If a switch is deployed within or between chassis, the switch may act as a translation between multiple protocols or technologies. When multiple chassis are connected to define a storage cluster, the storage cluster may be accessed by a client using either proprietary interfaces or standard interfaces such as network file system (NFS), common internet file system (CIFS), small computer system interface (SCSI) or hypertext transfer protocol (HTTP). Translation from the client protocol may occur at the switch, chassis external communication bus or within each storage node.

[0017] Each storage node may be one or more storage servers and each storage server is connected to one or more non-volatile solid-state memory units, which may be referred to as storage units. One embodiment includes a single storage server in each storage node and between one to eight non-volatile solid-state memory units, however this one example is not meant to be limiting. The storage server may include a processor, dynamic random access memory (DRAM) and interfaces for the internal communication bus and power distribution for each of the power buses. Inside the storage node, the interfaces and storage unit share a communication bus, e.g., PCI Express, in some embodiments. The non-volatile solid-state memory units may directly access the internal communication bus interface through a storage node communication bus, or request the storage node to access the bus interface. The non-volatile solid-state memory unit contains an embedded central processing unit (CPU), solid-state storage controller, and a quantity of solid-state mass storage, e.g., between 2-32 terabytes (TB) in some embodiments. An embedded volatile storage medium, such as DRAM, and an energy reserve apparatus are included in the non-volatile solid-state memory unit. In some embodiments, the energy reserve apparatus is a capacitor, super-capacitor, or battery that enables transferring a subset of DRAM contents to a stable storage medium in the case of power loss. In some embodiments, the non-volatile solid-state memory unit is constructed with a storage class memory, such as phase change memory (PCM) or other resistive random access memory (RRAM) or magnetoresistive random access memory (MRAM) that substitutes for DRAM and enables a reduced power hold-up apparatus.

[0018] One of many features of the storage nodes and non-volatile solid-state storage is the ability to proactively rebuild data in a storage cluster. The storage nodes and non-volatile solid-state storage can determine when a storage node or non-volatile solid-state storage in the storage cluster is unreachable, independent of whether there is an attempt to read data involving that storage node or non-volatile solid-state storage. The storage nodes and non-volatile solid-state storage then cooperate to recover and rebuild the data in at least partially new locations. In some embodiments the proactive rebuild enables the system to rebuild data without waiting until the data is needed for a read access initiated from a client communicating with the storage cluster.

[0019] With various combinations of the above features, and as further described below with reference to FIGS. 1-8, embodiments of a storage cluster recover from errors using error recovery mechanisms at multiple levels from a flash die, a storage, or a storage node, up to the multiple storage nodes that make up the storage cluster. A storage cluster may recover from an error by directing a flash die to apply error correction code at the die level, by retrying reads of the flash die with statistical probability analysis and/or with multiple values of a reference voltage or current, or by applying erasure coding during a read of user data across non-volatile solid-state storages. The storage cluster may recover from an error by reconfiguring storage nodes or non-volatile solid-state storages to a new erasure coding scheme, and rebuilding user data. The storage cluster could recover from loss of metadata, or loss of a remote procedure call cache, by accessing a redundant copy of the metadata or a redundant remote procedure call cache. In some embodiments, the storage cluster can determine an error is occurring during a read of user data, a non-volatile solid-state storage or storage node is unreachable, resources have been removed, lost, replaced or added, data has been lost or is otherwise unavailable, or metadata has been lost or is otherwise unavailable. The storage cluster may determine which of multiple recovery mechanisms to engage to recover from these situations as discussed in more detail below.

[0020] FIG. 1 is a perspective view of a storage cluster 160, with multiple storage nodes 150 and internal solid-state memory coupled to each storage node to provide network attached storage or storage area network, in accordance with some embodiments. A network attached storage, storage area network, or a storage cluster, or other storage memory, could include one or more storage clusters 160, each having one or more storage nodes 150, in a flexible and reconfigurable arrangement of both the physical components and the amount of storage memory provided thereby. The storage cluster 160 is designed to fit in a rack, and one or more racks can be set up and populated as desired for the storage memory. The storage cluster 160 has a chassis 138 having multiple slots 142. It should be appreciated that chassis 138 may be referred to as a housing, enclosure, or rack unit. In one embodiment, the chassis 138 has fourteen slots 142, although other numbers of slots are readily devised. For example, some embodiments have four slots, eight slots, sixteen slots, thirty-two slots, or other suitable number of slots. Each slot 142 can accommodate one storage node 150 in some embodiments. Chassis 138 includes flaps 148 that can be utilized to mount the chassis 138 on a rack. Fans 144 provide air circulation for cooling of the storage nodes 150 and components thereof, although other cooling components could be used, or an embodiment could be devised without cooling components. A switch fabric 146 couples storage nodes 150 within chassis 138 together and to a network for communication to the memory. In an embodiment depicted in FIG. 1, the slots 142 to the left of the switch fabric 146 and fans 144 are shown occupied by storage nodes 150, while the slots 142 to the right of the switch fabric 146 and fans 144 are empty and available for insertion of storage node 150 for illustrative purposes. This configuration is one example, and one or more storage nodes 150 could occupy the slots 142 in various further arrangements. The storage node arrangements need not be sequential or adjacent in some embodiments. Storage nodes 150 are hot pluggable, meaning that a storage node 150 can be inserted into a slot 142 in the chassis 138, or removed from a slot 142, without stopping or powering down the system. Upon insertion or removal of storage node 150 from slot 142, the system automatically reconfigures in order to recognize and adapt to the change. Reconfiguration, in some embodiments, includes restoring redundancy and/or rebalancing data or load.

[0021] Each storage node 150 can have multiple components. In the embodiment shown here, the storage node 150 includes a printed circuit board 158 populated by a CPU 156, i.e., processor, a memory 154 coupled to the CPU 156, and a non-volatile solid-state storage 152 coupled to the CPU 156, although other mountings and/or components could be used in further embodiments. The memory 154 has instructions which are executed by the CPU 156 and/or data operated on by the CPU 156. As further explained below, the non-volatile solid-state storage 152 includes flash or, in further embodiments, other types of solid-state memory.

[0022] Storage cluster 160 is scalable, meaning that storage capacity with non-uniform storage sizes is readily added, as described above. One or more storage nodes 150 can be plugged into or removed from each chassis and the storage cluster self-configures in some embodiments. Plug-in storage nodes 150, whether installed in a chassis as delivered or later added, can have different sizes. For example, in one embodiment a storage node 150 can have any multiple of 4 TB, e.g., 8 TB, 12 TB, 16 TB, 32 TB, etc. In further embodiments, a storage node 150 could have any multiple of other storage amounts or capacities. Storage capacity of each storage node 150 is broadcast, and influences decisions of how to stripe the data. For maximum storage efficiency, an embodiment can self-configure as wide as possible in the stripe, subject to a predetermined requirement of continued operation with loss of up to one, or up to two, non-volatile solid-state storage units 152 or storage nodes 150 within the chassis.

[0023] FIG. 2 is a block diagram showing a communications interconnect 170 and power distribution bus 172 coupling multiple storage nodes 150. Referring back to FIG. 1, the communications interconnect 170 can be included in or implemented with the switch fabric 146 in some embodiments. Where multiple storage clusters 160 occupy a rack, the communications interconnect 170 can be included in or implemented with a top of rack switch, in some embodiments. As illustrated in FIG. 2, storage cluster 160 is enclosed within a single chassis 138. External port 176 is coupled to storage nodes 150 through communications interconnect 170, while external port 174 is coupled directly to a storage node. External power port 178 is coupled to power distribution bus 172. Storage nodes 150 may include varying amounts and differing capacities of non-volatile solid-state storage 152. In addition, one or more storage nodes 150 may be a compute only storage node. Authorities 168 are implemented on the non-volatile solid-state storages 152, for example as lists or other data structures stored in memory. In some embodiments the authorities are stored within the non-volatile solid-state storage 152 and supported by software executing on a controller or other processor of the non-volatile solid-state storage 152. In a further embodiment, authorities 168 are implemented on the storage nodes 150, for example as lists or other data structures stored in the memory 154 and supported by software executing on the CPU 156 of the storage node 150. Authorities 168, which can be viewed as roles that the storage nodes 150 take on, control how and where data is stored in the non-volatile solid-state storages 152 in some embodiments. This control assists in determining which type of erasure coding scheme is applied to the data, and which storage nodes 150 have which portions of the data. Each authority 168 may be assigned to a non-volatile solid-state storage 152. Each authority may control a range of inode numbers, segment numbers, or other data identifiers which are assigned to data by a file system, by the storage nodes 150, or by the non-volatile solid-state storage 152, in various embodiments.

[0024] Every piece of data, and every piece of metadata, has redundancy in the system in some embodiments. In addition, every piece of data and every piece of metadata has an owner, which may be referred to as an authority. If that authority is unreachable, for example through failure of a storage node, there is a plan of succession for how to find that data or that metadata. In various embodiments, there are redundant copies of authorities 168. Authorities 168 have a relationship to storage nodes 150 and non-volatile solid-state storage 152 in some embodiments. Each authority 168, covering a range of data segment numbers or other identifiers of the data, may be assigned to a specific non-volatile solid-state storage 152. In some embodiments the authorities 168 for all of such ranges are distributed over the non-volatile solid-state storages 152 of a storage cluster. Each storage node 150 has a network port that provides access to the non-volatile solid-state storage(s) 152 of that storage node 150. Data can be stored in a segment, which is associated with a segment number and that segment number is an indirection for a configuration of a RAID (redundant array of independent disks) stripe in some embodiments. The assignment and use of the authorities 168 thus establishes an indirection to data. Indirection may be referred to as the ability to reference data indirectly, in this case via an authority 168, in accordance with some embodiments. A segment identifies a set of non-volatile solid-state storage 152 and a local identifier into the set of non-volatile solid-state storage 152 that may contain data. In some embodiments, the local identifier is an offset into the device and may be reused sequentially by multiple segments. In other embodiments the local identifier is unique for a specific segment and never reused. The offsets in the non-volatile solid-state storage 152 are applied to locating data for writing to or reading from the non-volatile solid-state storage 152 (in the form of a RAID stripe). Data is striped across multiple units of non-volatile solid-state storage 152, which may include or be different from the non-volatile solid-state storage 152 having the authority 168 for a particular data segment.

[0025] If there is a change in where a particular segment of data is located, e.g., during a data move or a data reconstruction, the authority 168 for that data segment should be consulted, at that non-volatile solid-state storage 152 or storage node 150 having that authority 168. In order to locate a particular piece of data, embodiments calculate a hash value for a data segment or apply an inode number or a data segment number. The output of this operation points to a non-volatile solid-state storage 152 having the authority 168 for that particular piece of data. This non-volatile solid-state storage 152, as the authority owner for the data segment, can coordinate a move or reconstruction of the data. In some embodiments there are two stages to this operation. The first stage maps an entity identifier (ID), e.g., a segment number, inode number, or directory number to an authority identifier. This mapping may include a calculation such as a hash or a bit mask. The second stage is mapping the authority identifier to a particular non-volatile solid-state storage 152, which may be done through an explicit mapping. The operation is repeatable, so that when the calculation is performed, the result of the calculation repeatably and reliably points to a particular non-volatile solid-state storage 152 having that authority 168. The operation may include the set of reachable storage nodes as input. If the set of reachable non-volatile solid-state storage units changes the optimal set changes. In some embodiments, the persisted value (i.e., the value that persistently results from the calculation) is the current assignment (which is always true) and the calculated value is the target assignment the cluster will attempt to reconfigure towards. This calculation may be used to determine the optimal non-volatile solid-state storage 152 for an authority in the presence of a set of non-volatile solid-state storage 152 that are reachable and constitute the same cluster. The calculation also determines an ordered set of peer non-volatile solid-state storages 152 that will also record the authority to non-volatile solid-state storage mapping so that the authority may be determined even if the assigned non-volatile solid-state storage is unreachable. A duplicate or substitute authority 168 may be consulted if a specific authority 168 is unavailable in some embodiments.

[0026] Two of the many tasks of the CPU 156 on a storage node 150 are to break up write data, and reassemble read data. When the system has determined that data is to be written, the authority 168 for that data is located as above. When the segment ID for data is-determined, the request to write is forwarded to the non-volatile solid-state storage 152 currently determined to be the host of the authority 168 determined from the segment. The host CPU 156 of the storage node 150, on which the non-volatile solid-state storage 152 and corresponding authority 168 reside, then breaks up or shards the data and transmits the data out to various non-volatile solid-state storages 152. In some embodiments, the authority 168 for the data segment being written to may defer sharding and distributing data to be done asynchronously after establishing redundancy for that data itself. The transmitted data is written as a data stripe in accordance with an erasure coding scheme. In some embodiments, data is requested to be pulled, and in other embodiments, data is pushed. In reverse, when data is read, the authority 168 for the segment ID containing the data is located as described above. The host CPU 156 of the storage node 150 on which the non-volatile solid-state storage 152 and corresponding authority 168 reside requests the data from the non-volatile solid-state storage and corresponding storage nodes pointed to by the authority. In some embodiments the data is read from flash storage as a data stripe. The host CPU 156 of storage node 150 then reassembles the read data, correcting any errors (if present) according to the appropriate erasure coding scheme, and forwards the reassembled data to the network. In further embodiments, some or all of these tasks can be handled in the non-volatile solid-state storage 152. In some embodiments, the segment host requests the data be sent to storage node 150 by requesting pages from storage and then sending the data to the storage node 150 making the original request. In some embodiments, a stripe width is only read if there is a single page read failure or delay.

[0027] In some systems, for example in UNIX-style file systems, data is handled with an index node or inode, which specifies a data structure that represents an object in a file system. The object could be a file or a directory, for example. Metadata may accompany the object, as attributes such as permission data and a creation timestamp, among other attributes. A segment number could be assigned to all or a portion of such an object in a file system. In other systems, data segments are handled with a segment number assigned elsewhere. For purposes of discussion, the unit of distribution is an entity, and an entity can be a file, a directory or a segment. That is, entities are units of data or metadata stored by a storage system. Entities are grouped into sets called authorities. Each authority has an authority owner, which is a storage node that has the exclusive right to update the entities in the authority. In other words, a storage node contains the authority, and that the authority, in turn, contains entities.

[0028] A segment is a logical container of data in accordance with some embodiments. A segment is an address space between medium address space and physical flash locations, i.e., the data segment number, are in this address space. Segments may also contain metadata, which enable data redundancy to be restored (rewritten to different flash locations or devices) without the involvement of higher level software. In one embodiment, an internal format of a segment contains client data and medium mappings to determine the position of that data. Each data segment is protected, e.g., from memory and other failures, by breaking the segment into a number of data and parity shards, where applicable. The data and parity shards are distributed, i.e., striped, across non-volatile solid-state storage 152 coupled to the host CPUs 156 in accordance with an erasure coding scheme. Usage of the term segments refers to the container and its place in the address space of segments in some embodiments. Usage of the term stripe refers to the same set of shards as a segment and includes how the shards are distributed along with redundancy or parity information in accordance with some embodiments.

[0029] A series of address-space transformations takes place across an entire storage system. At the top are the directory entries (file names) which link to an inode. Inodes point into medium address space, where data is logically stored. Medium addresses may be mapped through a series of indirect mediums to spread the load of large files, or implement data services like deduplication or snapshots. Medium addresses may be mapped through a series of indirect mediums to spread the load of large files, or implement data services like deduplication or snapshots. Segment addresses are then translated into physical flash locations. Physical flash locations have an address range bounded by the amount of flash in the system in accordance with some embodiments. Medium addresses and segment addresses are logical containers, and in some embodiments use a 128 bit or larger identifier so as to be practically infinite, with a likelihood of reuse calculated as longer than the expected life of the system. Addresses from logical containers are allocated in a hierarchical fashion in some embodiments. Initially, each non-volatile solid-state storage 152 may be assigned a range of address space. Within this assigned range, the non-volatile solid-state storage 152 is able to allocate addresses without synchronization with other non-volatile solid-state storage 152.

[0030] Data and metadata is stored by a set of underlying storage layouts that are optimized for varying workload patterns and storage devices. These layouts incorporate multiple redundancy schemes, compression formats and index algorithms. Some of these layouts store information about authorities and authority masters, while others store file metadata and file data. The redundancy schemes include error correction codes that tolerate corrupted bits within a single storage device (such as a NAND flash chip), erasure codes that tolerate the failure of multiple storage nodes, and replication schemes that tolerate data center or regional failures. In some embodiments, low density parity check (LDPC) code is used within a single storage unit. Reed-Solomon encoding is used within a storage cluster, and mirroring is used within a storage grid in some embodiments. Metadata may be stored using an ordered log structured index (such as a Log Structured Merge Tree), and large data may not be stored in a log structured layout.

[0031] In order to maintain consistency across multiple copies of an entity, the storage nodes agree implicitly on two things through calculations: (1) the authority that contains the entity, and (2) the storage node that contains the authority. The assignment of entities to authorities can be done by pseudorandomly assigning entities to authorities, by splitting entities into ranges based upon an externally produced key, or by placing a single entity into each authority. Examples of pseudorandom schemes are linear hashing and the Replication Under Scalable Hashing (RUSH) family of hashes, including Controlled Replication Under Scalable Hashing (CRUSH). In some embodiments, pseudo-random assignment is utilized only for assigning authorities to nodes because the set of nodes can change. The set of authorities cannot change so any subjective function may be applied in these embodiments. Some placement schemes automatically place authorities on storage nodes, while other placement schemes rely on an explicit mapping of authorities to storage nodes. In some embodiments, a pseudorandom scheme is utilized to map from each authority to a set of candidate authority owners. A pseudorandom data distribution function related to CRUSH may assign authorities to storage nodes and create a list of where the authorities are assigned. Each storage node has a copy of the pseudorandom data distribution function, and can arrive at the same calculation for distributing, and later finding or locating an authority. Each of the pseudorandom schemes requires the reachable set of storage nodes as input in some embodiments in order to conclude the same target nodes. Once an entity has been placed in an authority, the entity may be stored on physical devices so that no expected failure will lead to unexpected data loss. In some embodiments, rebalancing algorithms attempt to store the copies of all entities within an authority in the same layout and on the same set of machines.

[0032] Authority owners have the exclusive right to modify entities, to migrate entities from one non-volatile solid-state storage unit to another non-volatile solid-state storage unit, and to add and remove copies of entities. This allows for maintaining the redundancy of the underlying data. When an authority owner fails, is going to be decommissioned, or is overloaded, the authority is transferred to a new storage node. Transient failures make it non-trivial to ensure that all non-faulty machines agree upon the new authority location. The ambiguity that arises due to transient failures can be achieved automatically by a consensus protocol such as Paxos, hot-warm failover schemes, via manual intervention by a remote system administrator, or by a local hardware administrator (such as by physically removing the failed machine from the cluster, or pressing a button on the failed matching). In some embodiments, a consensus protocol is used, and failover is automatic. If too many failures or replication events occur in too short a time period, the system goes into a self-preservation mode and halts replication and data movement activities until an administrator intervenes in accordance with some embodiments.

[0033] As authorities are transferred between storage nodes and authority owners update entities in their authorities, the system transfers messages between the storage nodes and non-volatile solid-state storage units. With regard to persistent messages, messages that have different purposes are of different types. Depending on the type of the message, the system maintains different ordering and durability guarantees. As the persistent messages are being processed, the messages are temporarily stored in multiple durable and non-durable storage hardware technologies. In some embodiments, messages are stored in RAM, NVRAM and on NAND flash devices, and a variety of protocols are used in order to make efficient use of each storage medium. Latency-sensitive client requests may be persisted in replicated NVRAM, and then later NAND, while background rebalancing operations are persisted directly to NAND.

[0034] Persistent messages are persistently stored prior to being replicated. This allows the system to continue to serve client requests despite failures and component replacement. Although many hardware components contain unique identifiers that are visible to system administrators, manufacturers, hardware supply chains and ongoing monitoring quality control infrastructure, applications running on top of the infrastructure can address virtualized addresses. These virtualized addresses do not change over the lifetime of the storage system, regardless of component failures and replacements. This allows each component of the storage system to be replaced over time without reconfiguration or disruptions of client request processing.

[0035] In some embodiments, the virtualized addresses are stored with sufficient redundancy. A continuous monitoring system correlates hardware and software status and the hardware identifiers. This allows detection and prediction of failures due to faulty components and manufacturing details. The monitoring system also enables the proactive transfer of authorities and entities away from impacted devices before failure occurs by removing the component from the critical path in some embodiments.

[0036] In addition to component redundancy in the communication channel, storage cluster 160 is configured to allow for the loss of one or more storage nodes 150. In some embodiments this cluster redundancy level may be one for relatively small storage clusters 160 (less than 8 storage nodes 150) and two for relatively larger storage clusters 160 (8 or more storage nodes 150) although any number would be suitable for the cluster redundancy level. In some embodiments, where more storage nodes 150 than the redundancy level are lost, the storage cluster 160 cannot guarantee availability of data or integrity of future updates. As mentioned above, data redundancy is implemented via segments. A segment is formed by selecting equal sized shards from a subset of the non-volatile solid-state storage 152, each within a different storage node 150. Shards are reserved to establish the redundancy level, e.g., one or two, and then a remainder constitutes the data (the data shards). The shards are encoded using an ECC scheme such as parity or Reed-Soloman (RAID 6), so that any subset of the shards equal in count to the data shards may be used to reconstruct the complete data. The storage cluster redundancy represents a minimum level of redundancy and it may be exceeded for any individual data element. Segments are stored as a set of non-volatile solid-state storage units, roles (data position or parity) and allocation unit local to each non-volatile solid-state storage unit. The allocation units may be a physical address or an indirection determined within the non-volatile solid-state storage 152. Each shard may be portioned into pages and each page into code words. In some embodiments, the pages are between about 4 kilobytes (kB) and 64 kB, e.g., 16 kB, while the code words are between about 512 bytes to 4 kB, e.g., 1 kB. These sizes are one example and not meant to be limiting as any suitable size for the code words and the pages may be utilized. The code words contain local error correction and a checksum to verify the error correction was successful. This checksum is “salted” with the logical address of the contents meaning that a failure to match the checksum may occur if the data is uncorrectable or misplaced. In some embodiments, when a code word fails a checksum it is converted to an “erasure” for purpose of the error correction algorithm so that the code word may be rebuilt.

[0037] FIG. 3 is a multiple level block diagram, showing contents of a storage node 150 and contents of a non-volatile solid-state storage 152 of the storage node 150. Data is communicated to and from the storage node 150 by a network interface controller (NIC) 202 in some embodiments. Each storage node 150 has a CPU 156, and one or more non-volatile solid-state storage 152, as discussed above. Moving down one level in FIG. 3, each non-volatile solid-state storage 152 has a relatively fast non-volatile solid-state memory, such as non-volatile random access memory (NVRAM) 204, and flash memory 206. In some embodiments, NVRAM 204 may be a component that does not require program/erase cycles (DRAM, MRAM, PCM), and can be a memory that can support being written vastly more often than the memory is read from. Moving down another level in FIG. 3, the NVRAM 204 is implemented in one embodiment as high speed volatile memory, such as dynamic random access memory (DRAM) 216, backed up by energy reserve 218. Energy reserve 218 provides sufficient electrical power to keep the DRAM 216 powered long enough for contents to be transferred to the flash memory 206 in the event of power failure. In some embodiments, energy reserve 218 is a capacitor, super-capacitor, battery, or other device, that supplies a suitable supply of energy sufficient to enable the transfer of the contents of DRAM 216 to a stable storage medium in the case of power loss. The flash memory 206 is implemented as multiple flash dies 222, which may be referred to as packages of flash dies 222 or an array of flash dies 222. It should be appreciated that the flash dies 222 could be packaged in any number of ways, with a single die per package, multiple dies per package (i.e. multichip packages), in hybrid packages, as dies on a printed circuit board or other substrate. In some embodiments, the hybrid package may include a combination of memory types, such as NVRAM, random access memory (RAM), CPU, field programmable gate array (FPGA), or different sized flash memory in the same package. In the embodiment shown, the non-volatile solid-state storage 152 has a controller 212 or other processor, and an input output (I/O) port 210 coupled to the controller 212. I/O port 210 is coupled to the CPU 156 and/or the network interface controller 202 of the flash storage node 150. Flash input output (I/O) port 220 is coupled to the flash dies 222, and a direct memory access unit (DMA) 214 is coupled to the controller 212, the DRAM 216 and the flash dies 222. In the embodiment shown, the I/O port 210, controller 212, DMA unit 214 and flash I/O port 220 are implemented on a programmable logic device (PLD) 208, e.g., a field programmable gate array (FPGA). In this embodiment, each flash die 222 has pages, organized as sixteen kB (kilobyte) pages 224, and a register 226 through which data can be written to or read from the flash die 222. In further embodiments, other types of solid-state memory are used in place of, or in addition to flash memory illustrated within flash die 222.

[0038] In some embodiments, flash dies 222 have the ability to perform error correction, e.g., using error correction code, internal to the flash die 222. A non-volatile solid-state storage 152 having such a flash die 222 could recover from a read error internal to the flash die 222 by having the flash die 222 apply the error correction internally. In some embodiments, flash dies 222 have the ability to change a level at which a read is performed, for example by changing a reference voltage or a reference current internal to the flash die 222. The reference voltage or a reference current may be supplied external to the flash die 222. A non-volatile solid-state storage 152 having such a flash die could recover from a read error internal to the flash die 222 by having the flash die perform multiple retries, with multiple levels of the reference voltage or current. In some embodiments, a non-volatile solid-state storage 152 could perform multiple retries of a read from a flash die 222, with or without the above mentioned capabilities of the flash die 222, and perform probabilistic calculations as to whether the data is more likely to be a “1” or a “0”.

[0039] FIG. 4 is a block diagram showing a communication path or redundant copies of metadata, with further details of flash storage nodes and non-volatile solid-state storage units in accordance with some embodiments. Metadata 230 includes information about the user data that is written to or read from the flash memory 206. Metadata 230 can include messages, or derivations from the messages, indicating actions to be taken or actions that have taken place involving the data that is written to or read from the flash memory 206. Distributing redundant copies of metadata 230 to the non-volatile solid-state storage units 152 through the communications interconnect 170 ensures that messages are persisted and can survive various types of failure the system may experience. Each non-volatile solid-state storage 152 dedicates a portion of the NVRAM 204 to storing metadata 230. In many embodiments, redundant copies of metadata 230 are stored in the additional non-volatile solid-state storage 152.

[0040] Flash storage nodes 150 are coupled via the communication interconnect 170 (as described above with reference to FIG. 2). More specifically, the network interface controller 202 of each storage node 150 in the storage cluster is coupled to the communication interconnect 170, providing a communication path 234 among storage nodes 150 and non-volatile solid-state storage 152. Embodiments of storage nodes 150 have one or more non-volatile solid-state storage 152, as described above. Non-volatile solid-state storage 152 internal to a storage node can communicate with each other, for example via a bus, a serial communication path, a network path or other communication path 234 as readily devised in accordance with the embodiments disclosed herein.

[0041] Referring to FIGS. 3 and 4, in case of a power failure, whether local to non-volatile solid-state storage 152 or a storage node 150, data can be copied from the NVRAM 204 to the flash memory 206, if necessary, i.e., later versions of NVRAM may be persistent, such as PCM. For example, the DMA unit 214 (as seen in FIG. 3) can copy contents of the NVRAM 204, including the metadata, to the flash memory 206, using power supplied by the energy reserve 218. Energy reserve 218 (shown in FIG. 3) should be sized with sufficient capacity to support copy operation. That is energy reserve 218 should be sized so as to provide sufficient current at a sufficient voltage level for a time duration long enough to complete the copying. Messages that are in metadata 230 are persisted in the flash memory 206, in case of a power failure under this mechanism.

[0042] A further mechanism for persisting messages in a storage system involves the communication path 234 described above in FIG. 4. Redundant copies of the metadata 230 can be distributed via the communication path 234, in various ways. For example, a message coming from the filesystem could be distributed via the communication interconnect 170 as a broadcast over the communication path 234 to all of the non-volatile solid-state storage 152. A non-volatile solid-state storage 152 could send a copy of metadata 230 over the communication path 234 to other solid-non-volatile solid-state storage 152 in a storage node 150. In some embodiments, CPU 156 on a storage node 150, receiving a message from the communication interconnect 170 via the network interface controller 202 could send a copy of the message to each solid-state storage 152. The CPU 156 on a flash storage node 150 could rebroadcast the message to other flash storage nodes 150, and the flash storage nodes 150 could then distribute the message to the solid-state storages 152 in each of these flash storage nodes 150. In these and other uses of the communication path 234, redundant copies of the metadata 230 are distributed to the non-volatile solid-state storage 152. If one non-volatile solid-state storage unit 152, or one storage node 150 experiences a failure, redundant copies of any message of that non-volatile solid-state storage unit 152 or that storage node 150 are available in metadata 230 of other non-volatile solid-state storage 152. Each non-volatile solid-state storage 152 can apply decision logic 232 when evaluating various situations such as local power failure, an unreachable node, or a suggestion from a CPU 156, or from a controller 212, to consider or commence a data recovery or a data rebuild. The decision logic 232 includes witnessing logic, voting logic, consensus logic and/or other types of decision logic in various embodiments. Decision logic 232 could be implemented in hardware, software executing on the controller 212, firmware, or combinations thereof, and could be implemented as part of the controller 212 or coupled to the controller 212. The decision logic 232 is employed in consensus decisions among multiple solid-state storages 152, in some embodiments. In further embodiments, the decision logic 232 could cooperate with the other solid-state storages 152 in order to gather copies of the redundant metadata 230, and make local decisions. The mechanisms for persisting messages in a storage system are useful in the event of a failure, and can be used in data recovery and reconstruction as described above.

[0043] In some embodiments, flash memory can be written to once per location, until an entire block is erased, at which point flash memory can be written to again. If a failure occurs during a block erasure, or after a block erasure but prior to rewriting the flash memory, data corruption may result. Some actions span storage servers, such as writing data across a data stripe, and these messages use cluster messaging. If a failure occurs in the middle of such a data write, part of the stripe could be written and part of the stripe not written. Likewise, if a failure occurs in the middle of a data read, part of the stripe could be read and part of the stripe not read. If a file is created, and the file is in a directory, there are two records, one record is a directive and the other is the directory itself. In the instance of a failure, and only one of these has occurred, corruption results. The directive could indicate completion so that the file is created and stored, but the directory does not show the correct location or existence of the file. In some instances the directory may show the existence of a file in a directory tree, but there is no actual file in the storage.

[0044] Examples of messages include a request to write data, a request to read data, a request to lock or unlock a file, a change in permission of a file, an update to a file allocation table or other file or directory structure, a request to write a file that has executable instructions or to write a file name that is reserved and interpreted as an executable directory, updates to one or more authorities 168, updates to a fingerprint table, list or other data used in deduplication, updates to hash tables, updates to logs, and so on. When a message is received in non-volatile solid-state storage 152 of a storage node 150, indicating some action has taken place, the message or a derivation of the message is stored as metadata 230 in the NVRAM 204 of that solid-state storage 152. By applying the redundant copies of the metadata 230, actions are captured that are in progress, so that if a failure happens, these actions can be replayed and replacement actions can then be performed, for example upon restart. Actions span storage nodes and use cluster messaging, so the act of sending a message can be made persistent data via one or more of the mechanisms for persisting messages. In some embodiments, these messages don't require permanence beyond completion of the actions. In other embodiments these messages are further retained to facilitate rollback or other recovery operations.

[0045] For example, if a command is sent out to carry out a write operation, this message is recorded and redundant. If there is a failure, it can be determined whether or not that action has been carried out, and whether or not the action should be driven to completion. Such determination can be carried out using the decision logic 232 in each non-volatile solid-state storage 152. There is dedicated storage in NVRAM 204 for messages and other metadata 230, so that messages are recorded in the non-volatile solid-state storage 152 and replicated in some embodiments. The messages and other metadata 230 are written into flash memory 206 if one non-volatile solid-state storage 152 experiences a power failure, or if the entire system experiences a power failure or otherwise shuts down. The redundancy level of the messages thus matches the redundancy level of the metadata in some embodiments. When there are sufficient numbers of copies of messages, the message becomes irrevocable. If one node goes down, other nodes can vote, achieve consensus, or witness the various copies of the message and determine what action, if any, to carry to completion. If the entire system goes down, e.g., through a global power failure, then a sufficient number of these messages are written from NVRAM 204 to flash memory 206, so that upon restoration of power, the nodes can again open copies of the message and determine what action, if any, to carry to completion. It should be appreciated that the above mechanisms and actions prevent the sorts of corruptions described above.

[0046] Commands can come in the form of remote procedure calls, in various embodiments. A remote procedure call can be initiated by a client, and executed on one of the storage nodes 150 in the storage cluster 160. Some embodiments of the storage nodes 150 have redundant remote procedure call caches 354, 356, as described below with reference to FIG. 5. Remote procedure call caches offer fault tolerance that protects the system against loss of a remote procedure call in event of a failure. Some embodiments of the storage nodes 150 support multiple filesystems contemporaneously. Combining embodiments of the storage nodes 150 with redundant remote procedure call caches and the support for multiple filesystems offers fault-tolerant operation under multiple filesystems. It should be appreciated that a remote procedure call refers to a technology that allows a computer program to cause a subroutine or procedure to execute in another address space, commonly on another computing device on a shared network, without a programmer explicitly coding the details for this remote interaction in some embodiments.

[0047] In the above scenarios, the client could be operating over Internet Protocol, which may or may not be reliable. As an example, the client may retransmit a remote procedure call, but, meanwhile, the filesystem might have actually responded, which can lead to inconsistency (e.g. due to multiple executions of the same remote procedure call). A measure of fault tolerance, employing redundant remote procedure call caches can mitigate these potential problems, as described below.

[0048] FIG. 5 is a block diagram of storage nodes 150 with redundant remote procedure call caches 354, 356 in accordance with some embodiments. Each storage node 150 has an authority assignment table 352, a remote procedure call cache 354, and one or more mirrored remote procedure call caches 356. The remote procedure call cache 354 is located where the authority is located in order to minimize risk of breakage in a communication path between the location of the authority and the location of the remote procedure call cache 354 in some embodiments. Locating the remote procedure call cache 354 distant from the location of the authority is possible but may increase delays and increase risk of such breakage of a communication path.

[0049] In various embodiments, the remote procedure call cache 354 and mirrored remote procedure call cache 356 are implemented in the memory 154 coupled to the CPU 156 of a storage node 150 (see FIG. 1) or in the non-volatile random access memory 204 of the non-volatile solid-state storage 152 (see FIG. 3). In other embodiments, the remote procedure call cache 354 and mirrored remote procedure call cache 356 are implemented in the dynamic random access memory 216 coupled to the controller 212 in the non-volatile solid-state storage 152, in the flash 206 in the non-volatile solid-state storage 152, or in flash memory on a storage node 150. In one embodiment, the remote procedure call cache 354 and one or more mirrored remote procedure call caches 356 are implemented as metadata 230 in the non-volatile random access memory 204 of the non-volatile solid-state storage 152. In operation, a storage node 150 mirrors the remote procedure call cache 354 in at least one other storage node 150 of the storage cluster.

[0050] When a remote procedure call arrives for servicing, the storage node 150 or the non-volatile solid-state storage 152 determines whether the remote procedure call has already been serviced. This can be accomplished by checking the remote procedure call cache 354 to see if a result is already posted, i.e., the result of servicing the remote procedure call is available. For example, the result could be an acknowledgment that a data write or update to a directory structure has taken place, or the result could be error corrected data from a data read. If a result has been posted, the result is returned as a response to the remote procedure call, but the servicing internal to the storage node 150 or the non-volatile solid-state storage 152 is not repeated. In this manner, a repeated remote procedure call can be answered without causing inconsistency in the storage node 150 and/or non-volatile solid-state storage 152.

[0051] In a case where the remote procedure call cache 354 is unreachable, one or more of the remaining storage nodes 150 or non-volatile solid-state storages 152 locates the corresponding mirrored remote procedure call cache 356 and determines whether a result of servicing the remote procedure call is already posted. This scenario could occur, for example, if the non-volatile solid-state storage 152 or the storage node 150 having the remote procedure call cache 354 is unresponsive or otherwise unreachable. The result, if available from the mirrored remote procedure call cache 356, is then returned as above. If there is no result, from either the remote procedure call cache 354 or the mirrored remote procedure call cache 356 as appropriate, the remote procedure call is serviced and responded to with the result of that service.

[0052] In one embodiment, each storage node 150 mirrors the remote procedure call cache 354 in two other storage nodes 150, as depicted in FIG. 5. For example, the leftmost storage node 150 could send a copy of the contents of the remote procedure call cache 354 to two other storage nodes 150. Each of these other storage nodes 150 would place the copied contents of the remote procedure call cache 354 in a mirrored remote procedure call cache 356 of that storage node 150 as depicted by the arrows in FIG. 5. Each remote procedure call message has a unique transaction identifier, assigned by the client, uniquely identifying the transaction. If a storage node 150 is unreachable (whether permanently or temporarily), a copy of the remote procedure call cache 354 is available in at least one other storage node 150, e.g., in a mirrored remote procedure call cache 356. Each remote procedure call cache 354, and mirrored remote procedure call cache 356, contains the transaction identifier, the client identifier, and the result (e.g., an indication of whether or not the action is complete), in one embodiment. The remote procedure call and information relating thereto as stored in the remote procedure call cache 354, are forms of metadata in some embodiments.

[0053] In one embodiment, each storage node 150 consults the table 352 resident in that storage node 150. For example, table 352 could reside in the memory of the storage node 150, or the memory of a non-volatile solid-state storage 152 of the storage node 150, and so on. The remote procedure call cache 354 is maintained for an authority for which that storage node 150 has primary authority, as indicated in table 352. The storage node 150 mirrors the remote procedure call cache 354, by sending updates (copies) of the contents of the remote procedure call cache 354 to the storage node 150 identified as having the first backup authority, and to the storage node 150 identified as having the second backup authority, according to the table 352. In further embodiments, additional copies of the remote procedure call cache 354 could be mirrored, or the copies and mirror caches could be distributed in a different manner. Locations of the mirrored remote procedure call caches 356 could be tracked by another table or tracking mechanism instead of table 352 in other embodiments.

[0054] In a scenario where a mirrored remote procedure call cache 356 belonging to a non-volatile solid-state storage 152 or a storage node 150 is unreachable (e.g., if the solid-state storage 152 or non-volatile storage node 150 itself is unreachable), the remaining storage nodes 150 can determine and assign a replacement mirrored remote procedure call cache 356. The determination and assignment of a replacement mirrored remote procedure call cache 356 may include applying one or more mechanisms, such as witnessing, voting, volunteering, consulting the table 352 to find a backup authority and assigning the mirrored remote procedure call cache 356 to the same node as has the backup authority, and so on. In some embodiments, a mirrored remote procedure call cache 356 could be assigned to a differing node than the node having the backup authority. Once the replacement mirrored remote procedure call cache 356 is determined, the storage node 150 corresponding to the replacement mirrored remote procedure call cache 356 can mirror the remote procedure call cache 354 to the mirrored remote procedure call cache 356. The remote procedure call may then be serviced and responded to with a result. In some embodiments, remote procedure call entries to the remote procedure call cache 354 and/or the mirrored remote procedure call cache(s) 356 are deleted after a time expires, e.g., through the use of one or more timers, or timestamps, etc.

[0055] The ability to change configuration of data stripes 314 is another one of many features of storage nodes 150 and non-volatile solid-state storages 152. In the examples of FIGS. 6A and 6B data is stored in the form of data stripes 314, 326, 338, 340, 342, 344, in which data is sharded, i.e., broken up and distributed, across multiple storage nodes 150. In some embodiments, data can be striped across non-volatile solid-state storages 152 in a storage node 150 or in multiple storage nodes 150. Various RAID configurations and associated data striping schemes and levels of redundancy are possible, as controlled by the authority 168 for each data segment. A storage node 150 having an authority 168 for a specified data segment could be a storage node 150 acting as a data server, a storage node 150 acting as a parity server, or a storage node 150 having no solid-state storage 152 (compute only). In various arrangements, each storage node 150, or each non-volatile solid-state storage 152, provides one bit of data or one parity bit, for a data stripe 314. Various embodiments implement an error correction code (ECC) that allows recovery of data even if one or two storage nodes 150 fail or are unavailable. In some embodiments, the storage node holding the authority 168 determines which RAID scheme or level and authority 168 points to the data stripe 314. Data striping can be applied at a bit level, byte level or block level, and further data striping schemes are possible. In some versions, a storage node 150, or a solid-state storage 152, could contribute more than one bit to a data stripe 314. In some embodiments, an error correction code calculation is performed at each non-volatile solid-state storage 152, for the shard of data. That shard of data, corrected as needed, is sent back to the storage node 150 that has the authority 168 for the data, where the data is reassembled. More than one stripe type, or RAID scheme or level, can be present across the storage nodes 150 (i.e., coexist in the storage cluster 160), in some embodiments.

[0056] FIG. 6A is a configuration diagram of data stripes 314, 326 of differing sizes, i.e., differing stripe widths. These data stripes 314, 326 coexist across the storage nodes 150, or in some embodiments coexist across non-volatile solid-state storage 152. For example, one data stripe 326 is sharded across the three storage nodes 150 in a RAID scheme or level having double redundancy. Identical copies of the data from a first storage node 150 are present in each of the second storage node 150 and the third storage node 150. Since this version of data recovery, in the example data stripe 326, requires two identical copies of data, the storage overhead is 200% more than the data storage capacity. In other words, the relative total storage amount is (N+2N) divided by N, which equals three, for N bits of data, and this is independent of N. Another data stripe 314 is sharded across storage nodes 150 acting as data servers and across storage nodes 150 acting as parity servers providing the p and q parity bits. User data is written to and read from the storage nodes 150 in accordance with a RAID scheme or level having two parity bits, i.e., parity bit p and parity bit q, in this data stripe 314. Since this particular error correction code adds two bits to the data length, the relative storage overhead is related to (N+2) divided by N, for N bits of data. For example, 10 plus 2 redundancy has 20% overhead of memory. The wider data stripe 314 therefore has greater storage efficiency and lower storage overhead than the narrower data stripe 326.

[0057] FIG. 6B is a configuration diagram of data stripes 338, 340, 342, 344 across storage nodes 150 of various memory capacities in accordance with some embodiments. As illustrated, two of the storage nodes 150 have greater capacity than two others of the storage nodes 150, for example by a factor of two. All of the capacity of these storage nodes 150 can be used by applying data stripes 338, 340, 342, 344 as shown. For example, two data stripes 338, 340 are applied across two of the higher capacity storage nodes 150 and one of the lower capacity storage nodes 150. Two more data stripes 342, 344 are applied across one of the lower capacity storage nodes 150 and two of the higher capacity storage nodes 150.

[0058] The ability of various embodiments to self-configure, on power up or upon removal, replacement or insertion of one or more storage nodes 150 and/or solid-state storages 152 provides a storage memory topology that automatically reconfigures. For example, in a storage cluster with multiple storage nodes 150 and two levels of redundancy, two storage nodes 150 could be lost and data could still be reconstructed, as described above. In a small configuration, the storage cluster 160 could self-configure to store two replicated copies, i.e., mirrors of the data, with 200% storage overhead. In a larger configuration, the cluster could self-configure to have parity pages, with a lower storage overhead. Storage overhead is thus reconfigured as cluster membership changes. The storage nodes 150, solid-state storages 152, and storage cluster 160 which these form can dynamically switch between RAID schemes, and at any moment could have a hybrid combination of RAID schemes. Earlier-formed stripes do not need to be reconstructed when the topology of the storage cluster 160 changes and can be left as is, or reconstructed later according to a new RAID scheme. Storage nodes 150, and non-volatile solid-state storage 152 can switch from one data striping scheme to another in subsequent accesses, i.e., writes or reads, in some embodiments. New data that is arriving can be written to whichever topology is in place at the moment of the data arrival. Adding one or more storage nodes 150, or non-volatile solid-state storage 152, does not require that data be removed from the system for system repartitioning. The topology of the storage cluster 160, e.g., the RAID scheme(s) and storage overhead, are automatically reconfigured as the geometry of the storage cluster 160 and/or storage capacity of each storage node 150 or each non-volatile solid-state storage 152, is changed. The storage nodes 150 and non-volatile solid-state storage 152 thus implement dynamically switching between data striping, e.g., RAID, schemes in a hybrid topology of storage memory.

[0059] FIG. 7 is a flow diagram of a method for error recovery in a storage cluster, which can be practiced on or by embodiments of the storage cluster, storage nodes and/or solid-state storages disclosed herein in accordance with some embodiments. The storage cluster writes user data, using erasure coding, across the storage nodes and the non-volatile solid-state storages of the storage cluster, in an action 702. This can be performed by the various processors and modules at the various levels of the storage cluster. In an action 704, metadata and a redundant copy or copies of the metadata are written across the storage nodes and non-volatile solid-state storages of the storage cluster. Data reads, further data writes, and further accesses to the metadata along with other operations are performed in an action 706, which may occur during normal system operation of the storage cluster. In a decision action 708 it is determined if an error is detected. If no error is detected, flow branches back to the action 706 for further system operation. If an error is detected, flow proceeds to action 710.

[0060] In the action 710, the storage cluster (i.e., the storage nodes and the non-volatile solid-state storages of the storage cluster) determines whether to recover from the error at the die level, i.e., at the level of the flash die, dies or dice, or package, the non-volatile solid-state storage or controller level, the storage node or processor level, and/or the cluster level, i.e., at the level of the plurality of storage nodes. In an action 712, the recovery mechanism or mechanisms at the determined level or levels is engaged. With the recovery in progress and/or completed, flow then branches back to the action 706, for further system operation as described above. Examples are provided below that summarize various errors and engagement of various recovery mechanisms at various levels in various embodiments, as described in detail with reference to FIGS. 1-6. The examples provided are illustrative and not meant to be an exhaustive list as other scenarios can be implemented with the embodiments described herein.

• • [0061] An error in reading the flash memory could lead to engagement of an error recovery at the level of the flash memory such as the flash memory applying error correction to data stored in the flash memory. Such an error could lead to engagement of an error recovery at the level of the non-volatile solid-state storage or controller in the non-volatile solid-state storage such as retrying the read of the flash memory. This could involve applying multiple levels of a reference voltage or reference current in successive read retries, and/or probabilistic calculation or other statistical analysis to the result of the read retries.

• • [0062] An error in reading user data could lead to engagement of an error recovery at the level of the storage node such as applying erasure coding to data shards read from the storage nodes. An error in reading user data, or an error from loss of one or two storage nodes could lead to an error recovery at the level of the cluster or storage nodes, such as rebuilding the user data onto whichever storage nodes remain in service. This could involve a newly assigned erasure coding for rewriting the recovered data, which differs from the erasure coding by which the data was earlier written and by which the data is read during the recovery of the data.

• • [0063] An error from loss of the owner of a data segment, e.g., when an initial attempt to read data fails and/or one of the storage nodes or non-volatile solid-state storages that owns the data segment is unreachable, could lead to an error recovery at the level of the cluster or storage nodes. This could include reassigning ownership of a data segment to another one of the storage nodes or non-volatile solid-state storages. This could further lead to recovering and rebuilding the data.

• • [0064] An error from loss of metadata could lead to locating and accessing a redundant copy of the metadata. This loss of metadata could be a result of a failure of a memory holding the metadata, or a failure of a non-volatile solid-state storage or a storage node. The redundant copy of the metadata may be located on another one of the storage nodes or non-volatile solid-state storages. This type of error recovery engages an error recovery mechanism at the level of the storage nodes or the non-volatile solid-state storages.

• • [0065] An error from loss of a remote procedure call cache could lead to locating and accessing a redundant copy of the remote procedure call cache, i.e., a mirrored remote procedure call cache. Similar to the loss of metadata, this could be a result of a failure of a memory holding the remote procedure call cache, or a failure of a non-volatile solid-state storage or a storage node. The mirrored remote procedure call cache could be located on another one of the storage nodes or non-volatile solid-state storages. This type of error recovery engages an error recovery mechanism at the level of the storage nodes or the non-volatile solid-state storages.

[0066] It should be appreciated that the methods described herein may be performed with a digital processing system, such as a conventional, general-purpose computer system. Special purpose computers, which are designed or programmed to perform only one function may be used in the alternative. FIG. 8 is an illustration showing an exemplary computing device which may implement the embodiments described herein. The computing device of FIG. 8 may be used to perform embodiments of the functionality for a storage node or a non-volatile solid-state storage in accordance with some embodiments. The computing device includes a central processing unit (CPU) 801, which is coupled through a bus 805 to a memory 803, and mass storage device 807. Mass storage device 807 represents a persistent data storage device such as a disc drive, which may be local or remote in some embodiments. The mass storage device 807 could implement a backup storage, in some embodiments. Memory 803 may include read only memory, random access memory, etc. Applications resident on the computing device may be stored on or accessed via a computer readable medium such as memory 803 or mass storage device 807 in some embodiments. Applications may also be in the form of modulated electronic signals modulated accessed via a network modem or other network interface of the computing device. It should be appreciated that CPU 801 may be embodied in a general-purpose processor, a special purpose processor, or a specially programmed logic device in some embodiments.

[0067] Display 811 is in communication with CPU 801, memory 803, and mass storage device 807, through bus 805. Display 811 is configured to display any visualization tools or reports associated with the system described herein. Input/output device 809 is coupled to bus 805 in order to communicate information in command selections to CPU 801. It should be appreciated that data to and from external devices may be communicated through the input/output device 809. CPU 801 can be defined to execute the functionality described herein to enable the functionality described with reference to FIGS. 1-7. The code embodying this functionality may be stored within memory 803 or mass storage device 807 for execution by a processor such as CPU 801 in some embodiments. The operating system on the computing device may be MS-WINDOWS™, UNIX™, LINUX™, iOS™, CentOS™, Android™, Redhat Linux™, z/OS™, or other known operating systems. It should be appreciated that the embodiments described herein may be integrated with virtualized computing system also.

[0068] Detailed illustrative embodiments are disclosed herein. However, specific functional details disclosed herein are merely representative for purposes of describing embodiments. Embodiments may, however, be embodied in many alternate forms and should not be construed as limited to only the embodiments set forth herein.

[0069] It should be understood that although the terms first, second, etc. may be used herein to describe various steps or calculations, these steps or calculations should not be limited by these terms. These terms are only used to distinguish one step or calculation from another. For example, a first calculation could be termed a second calculation, and, similarly, a second step could be termed a first step, without departing from the scope of this disclosure. As used herein, the term “and/or” and the “/” symbol includes any and all combinations of one or more of the associated listed items.

[0070] As used herein, the singular forms “a”, “an” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms “comprises”, “comprising”, “includes”, and/or “including”, when used herein, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof. Therefore, the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting.

[0071] It should also be noted that in some alternative implementations, the functions/acts noted may occur out of the order noted in the figures. For example, two figures shown in succession may in fact be executed substantially concurrently or may sometimes be executed in the reverse order, depending upon the functionality/acts involved.

[0072] With the above embodiments in mind, it should be understood that the embodiments might employ various computer-implemented operations involving data stored in computer systems. These operations are those requiring physical manipulation of physical quantities. Usually, though not necessarily, these quantities take the form of electrical or magnetic signals capable of being stored, transferred, combined, compared, and otherwise manipulated. Further, the manipulations performed are often referred to in terms, such as producing, identifying, determining, or comparing. Any of the operations described herein that form part of the embodiments are useful machine operations. The embodiments also relate to a device or an apparatus for performing these operations. The apparatus can be specially constructed for the required purpose, or the apparatus can be a general-purpose computer selectively activated or configured by a computer program stored in the computer. In particular, various general-purpose machines can be used with computer programs written in accordance with the teachings herein, or it may be more convenient to construct a more specialized apparatus to perform the required operations.

[0073] A module, an application, a layer, an agent or other method-operable entity could be implemented as hardware, firmware, or a processor executing software, or combinations thereof. It should be appreciated that, where a software-based embodiment is disclosed herein, the software can be embodied in a physical machine such as a controller. For example, a controller could include a first module and a second module. A controller could be configured to perform various actions, e.g., of a method, an application, a layer or an agent.

[0074] The embodiments can also be embodied as computer readable code on a non-transitory computer readable medium. The computer readable medium is any data storage device that can store data, which can be thereafter read by a computer system. Examples of the computer readable medium include hard drives, network attached storage (NAS), read-only memory, random-access memory, CD-ROMs, CD-Rs, CD-RWs, magnetic tapes, and other optical and non-optical data storage devices. The computer readable medium can also be distributed over a network coupled computer system so that the computer readable code is stored and executed in a distributed fashion. Embodiments described herein may be practiced with various computer system configurations including hand-held devices, tablets, microprocessor systems, microprocessor-based or programmable consumer electronics, minicomputers, mainframe computers and the like. The embodiments can also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a wire-based or wireless network.

[0075] Although the method operations were described in a specific order, it should be understood that other operations may be performed in between described operations, described operations may be adjusted so that they occur at slightly different times or the described operations may be distributed in a system which allows the occurrence of the processing operations at various intervals associated with the processing.

[0076] In various embodiments, one or more portions of the methods and mechanisms described herein may form part of a cloud-computing environment. In such embodiments, resources may be provided over the Internet as services according to one or more various models. Such models may include Infrastructure as a Service (IaaS), Platform as a Service (PaaS), and Software as a Service (SaaS). In IaaS, computer infrastructure is delivered as a service. In such a case, the computing equipment is generally owned and operated by the service provider. In the PaaS model, software tools and underlying equipment used by developers to develop software solutions may be provided as a service and hosted by the service provider. SaaS typically includes a service provider licensing software as a service on demand. The service provider may host the software, or may deploy the software to a customer for a given period of time. Numerous combinations of the above models are possible and are contemplated.

[0077] Various units, circuits, or other components may be described or claimed as “configured to” perform a task or tasks. In such contexts, the phrase “configured to” is used to connote structure by indicating that the units/circuits/components include structure (e.g., circuitry) that performs the task or tasks during operation. As such, the unit/circuit/component can be said to be configured to perform the task even when the specified unit/circuit/component is not currently operational (e.g., is not on). The units/circuits/components used with the “configured to” language include hardware—for example, circuits, memory storing program instructions executable to implement the operation, etc. Reciting that a unit/circuit/component is “configured to” perform one or more tasks is expressly intended not to invoke 35 U.S.C. 112, sixth paragraph, for that unit/circuit/component. Additionally, “configured to” can include generic structure (e.g., generic circuitry) that is manipulated by software and/or firmware (e.g., an FPGA or a general-purpose processor executing software) to operate in manner that is capable of performing the task(s) at issue. “Configured to” may also include adapting a manufacturing process (e.g., a semiconductor fabrication facility) to fabricate devices (e.g., integrated circuits) that are adapted to implement or perform one or more tasks.

[0078] The foregoing description, for the purpose of explanation, has been described with reference to specific embodiments. However, the illustrative discussions above are not intended to be exhaustive or to limit the invention to the precise forms disclosed. Many modifications and variations are possible in view of the above teachings. The embodiments were chosen and described in order to best explain the principles of the embodiments and its practical applications, to thereby enable others skilled in the art to best utilize the embodiments and various modifications as may be suited to the particular use contemplated. Accordingly, the present embodiments are to be considered as illustrative and not restrictive, and the invention is not to be limited to the details given herein, but may be modified within the scope and equivalents of the appended claims.