Error detection/correction and fault detection/recovery – Data processing system error or fault handling – Reliability and availability
Reexamination Certificate
2000-10-10
2004-02-10
Iqbal, Nadeem (Department: 2184)
Error detection/correction and fault detection/recovery
Data processing system error or fault handling
Reliability and availability
C714S015000
Reexamination Certificate
active
06691245
ABSTRACT:
FIELD OF THE INVENTION
This invention relates to data storage in a computerized network or system. More particularly, the present invention relates to a new and improved technique of host-initiated synchronization of data that is stored on both a local storage device and a remote mirroring fail-over storage device. The data stored by the host on the local storage device is mirrored to the remote storage device, and a synchronization procedure enables the host and remote storage device easily and quickly to “roll back” to, and continue operations from, a stable, coherent state in the event of a failure of the local storage device.
BACKGROUND OF THE INVENTION
Computerized systems are commonly used to operate various businesses or enterprises. In many cases, the data that is kept on the computers and data storage devices is critical to the functioning of the enterprise. A temporary inability to access this data can halt business operations, and a total loss or corruption of the data can severely cripple the entire enterprise. Therefore, it is important to such enterprises to maintain availability and validity of the data.
One technique to ensure data availability and validity is to store the data in more than one storage device, such as in primary and secondary storage devices. In this case, the secondary storage device maintains a “mirrored,” or duplicate, copy of the data. In the event of a failure of the primary storage device, operations can resume using the secondary storage device and the mirrored data.
Additionally, the secondary storage device is typically maintained at a geographically remote location from the primary storage device, such as at a different city or state, while the primary storage device is kept locally. In this manner, a geographical disturbance, such as a local citywide power outage, will not affect both storage devices, and operations can eventually resume.
Also, the local and remote storage devices are typically accessed by host devices, or storage servers, that serve the data storage requirements of various client devices. At least one such host device is maintained at the local site and another at the remote location to access the local and remote storage devices, respectively. Therefore, when the local storage device fails, the remote host device, using the remote storage device, takes over serving the data storage requirements of the various clients.
Various methods have been developed to mirror, or duplicate, the data from the primary storage device at the local site to the alternate, secondary storage device at the remote site. Such remote mirroring solutions ensure the continuance of business in the event of a geographical disaster. Many of these solutions, however, have either performance or coherency synchronization issues. Performance issues require that very little time be taken to perform a “fail-over” to, or switch to, the remote storage and host devices, so as not to degrade the overall performance of the clients using the backed-up data. Coherency synchronization requires that the state of the stored data between the local and remote storage devices, be put in a “coherent state” at which it is assured that both have correct, up-to-date data that may be used by a file system or database. In the event of a fail-over situation, the synchronization difficulties of current mirroring techniques can result in time-consuming special efforts to generate a coherent state in the remote storage device through file system check and recovery procedures, so that applications executing on the various clients can proceed to operate.
It is with respect to these and other background considerations that the present invention has evolved.
SUMMARY OF THE INVENTION
The present invention enables efficient remote data mirroring and “fail-over” capabilities in a computer system wherein a local host device stores data on a local storage device on behalf of various client devices, and mirrors the data storage on a remote storage device. “Fail-over” refers to a situation in which the local storage device can no longer service data access requests, so the client devices must switch to using remote storage device with a remote host device for data backup processing. The local host device periodically initiates data synchronization procedures for the local and remote storage devices. Information regarding the latest synchronization procedures is maintained within the local host, local storage and remote storage devices. The synchronization information defines a common, known, coherent state of stored data for all of these devices. The time at which a data synchronization occurs is called a “checkpoint,” and the condition of the stored data at which the coherent state is defined is called the “checkpoint state.”
The remote storage device maintains a “snapshot” of the data at the latest checkpoint state. The snapshot is essentially a copy of a portion of the data as the data existed at the last checkpoint state. Changes to the stored data on the remote storage device are accepted after each previously occurring checkpoint, but the data that was present at the last checkpoint is transferred to and preserved in the snapshot, so it can be restored at a later time if necessary.
Since the remote storage device maintains information describing the checkpoint state, in the event of a fail-over condition, the remote host device quickly and easily “rolls back” the state of the data stored on the remote storage device to the last common checkpoint state. The data is restored from the snapshot. Applications executing on the client devices, thus, restart at the restored checkpoint state with a minimum of interruption.
These and other improvements are achieved by storing and synchronizing data between a host device, a primary storage device and a secondary storage device. The host device stores data on the primary storage device on behalf of client devices. The data stored on the primary storage device is mirrored to the secondary storage device. Data synchronization between the host device and the primary storage device is initiated by the host device. A checkpoint message is issued from the host device to the primary storage device. The checkpoint message indicates that a storage state of the host device is at a stable consistent state. Data synchronization between the primary and secondary storage devices is performed by the primary storage device. The checkpoint message is then forwarded from the primary storage device to the secondary storage device. An incremental snapshot of the mirrored data is generated on the secondary storage device at the predetermined checkpoint indicated by the checkpoint message. The incremental snapshot includes data and information describing the mirrored data at the predetermined checkpoint to preserve a storage state of the secondary storage device at the predetermined checkpoint.
It is preferable that data be sent from the host device to the primary storage device and forwarded to the secondary storage device, so both the primary and secondary storage devices can update their storage state to be consistent with the host device. It is further preferable, when new data is sent from the host device to the primary storage device and then to the secondary storage device after the predetermined checkpoint, that the secondary storage device transfer any preexisting data, if it is replaced by the new data, to the incremental snapshot. Thus, the incremental snapshot maintains the storage state of the secondary storage device at the predetermined checkpoint.
It is also preferable that these steps be performed in conjunction with failing-over from utilization of the first host device and the primary storage device to utilization of a second host device and the secondary storage device. In such a fail-over situation, a failure of the first host device and/or the primary storage device is detected, and the second host device and the secondary storage device are signaled that they are to be utilized for primary data storage. An image of the data stored on the
Bonura Timothy M.
Iqbal Nadeem
LSI Logic Corporation
Suiter - West PC LLO
LandOfFree
Data storage with host-initiated synchronization and... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Data storage with host-initiated synchronization and..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Data storage with host-initiated synchronization and... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3342623