公开(公告)号：US20060101033A1

公开(公告)日：2006-05-11

申请号：US10982135

申请日：2004-11-05

申请人： Wei Hu ,  Yunrui Li ,  Vinay Srihari ,  Ramana Yerneni

发明人： Wei Hu ,  Yunrui Li ,  Vinay Srihari ,  Ramana Yerneni

IPC分类号： G06F7/00

CPC分类号： G06F9/46 ,  G06F11/1464 ,  G06F11/1471 ,  G06F11/2035 ,  G06F11/2046 ,  G06F11/2097 ,  G06F2201/80

摘要： Each of a plurality of Worker processes are allowed to perform any and all of the following tasks involving logged work items: (1) reading a subset of the work items from a log; (2) sequentially ordering work items for corresponding data objects; (3) applying a sequentially ordered set of work items to a corresponding data object; and (4) transmitting a subset of work items to a Worker process running on another database server in a cluster, if necessary. These tasks can be performed concurrently, at will, and as available, by the Worker processes. An improved checkpointing technique eliminates the need for the Worker processes to get to a synchronization point and stop. Instead, a Coordinator process examines the current state of progress of the Worker processes and computes a past point in the sequence of work items at which all work items before that point have been completely processed, and records this point as the checkpoint.

公开(公告)号：US07020599B1

公开(公告)日：2006-03-28

申请号：US10017337

申请日：2001-12-13

申请人： Qiang Cao ,  Sushil Kumar ,  Tirthankar Lahiri ,  Yunrui Li ,  Gianfranco Putzolu

发明人： Qiang Cao ,  Sushil Kumar ,  Tirthankar Lahiri ,  Yunrui Li ,  Gianfranco Putzolu

IPC分类号： G06F9/45

CPC分类号： G06F11/3457 ,  G06F17/30306 ,  G06F2201/86 ,  Y10S707/99943 ,  Y10S707/99953

摘要： A method for simulating different MTTR settings includes determining a simulated MTTR setting and providing a simulated checkpoint queue. The simulated checkpoint queue is associated with the simulated MTTR setting and is an ordered list of one or more elements. Each element represents a buffer, and the ordered list has a head and a tail. The method also includes providing a simulated write counter associated with the simulated MTTR setting. The method further includes, in response to detecting a change to a first buffer, checking if the first buffer is represented in the simulated checkpoint queue. If the first buffer is not represented in the simulated checkpoint queue, an element that represents the first buffer is linked to the tail of the simulated checkpoint queue. An MTTR advisory system includes a memory, one or more processors coupled to the memory, a simulated MTTR setting, a simulated checkpoint queue, and a simulated write counter. The simulated MTTR setting is maintained in the memory. The simulated checkpoint queue is maintained in the memory and associated with the simulated MTTR setting. The simulated write counter is also maintained in the memory, and is associated with the simulated MTTR setting. The simulated write counter provides a count of the number of times an element is removed from the simulated checkpoint queue, wherein the element is removed from the simulated checkpoint queue in response to a write out of a buffer from volatile memory and storing in nonvolatile memory.

摘要翻译： 用于模拟不同MTTR设置的方法包括确定模拟MTTR设置并提供模拟检查点队列。 模拟的检查点队列与模拟的MTTR设置相关联，并且是一个或多个元素的有序列表。 每个元素表示一个缓冲区，有序列表有一个头和尾。 该方法还包括提供与模拟的MTTR设置相关联的模拟写入计数器。 该方法还包括响应于检测到对第一缓冲器的改变，检查在模拟检查点队列中是否表示第一缓冲器。 如果模拟检查点队列中没有表示第一个缓冲区，则表示第一个缓冲区的元素链接到模拟检查点队列的尾部。 MTTR咨询系统包括存储器，耦合到存储器的一个或多个处理器，模拟MTTR设置，模拟检查点队列和模拟写入计数器。 模拟的MTTR设置保留在内存中。 模拟检查点队列保存在存储器中并与模拟的MTTR设置相关联。 模拟写计数器也保存在存储器中，并与模拟的MTTR设置相关联。 仿真写入计数器提供从模拟检查点队列中删除元素的次数的计数，其中响应于从易失性存储器写入缓冲器并存储在非易失性存储器中，将元件从模拟检查点队列中移除。

公开(公告)号：US08566326B2

公开(公告)日：2013-10-22

申请号：US10982135

申请日：2004-11-05

申请人： Wei Hu ,  Yunrui Li ,  Vinay Srihari ,  Ramana Yemeni

发明人： Wei Hu ,  Yunrui Li ,  Vinay Srihari ,  Ramana Yemeni

IPC分类号： G06F17/30

CPC分类号： G06F9/46 ,  G06F11/1464 ,  G06F11/1471 ,  G06F11/2035 ,  G06F11/2046 ,  G06F11/2097 ,  G06F2201/80

摘要： Each of a plurality of Worker processes are allowed to perform any and all of the following tasks involving logged work items: (1) reading a subset of the work items from a log; (2) sequentially ordering work items for corresponding data objects; (3) applying a sequentially ordered set of work items to a corresponding data object; and (4) transmitting a subset of work items to a Worker process running on another database server in a cluster, if necessary. These tasks can be performed concurrently, at will, and as available, by the Worker processes. An improved checkpointing technique eliminates the need for the Worker processes to get to a synchronization point and stop. Instead, a Coordinator process examines the current state of progress of the Worker processes and computes a past point in the sequence of work items at which all work items before that point have been completely processed, and records this point as the checkpoint.

摘要翻译： 允许多个Worker进程中的每一个执行涉及记录的工作项的任何和所有以下任务：（1）从日志读取工作项的子集; （2）顺序排列相应数据对象的工作项; （3）将顺序排列的工作项集合应用于对应的数据对象; （4）如果需要，将工作项的子集发送到在群集中的另一数据库服务器上运行的工作进程。 这些任务可以在工作进程的同时，随意和可用的情况下执行。 改进的检查点技术消除了Worker进程到达同步点并停止的需要。 相反，协调器进程将检查工作进程的当前状态，并计算工作项序列中的过去点，在此之前，该点之前的所有工作项都已完全处理，并将此点记录为检查点。

公开(公告)号：US08615578B2

公开(公告)日：2013-12-24

申请号：US11245868

申请日：2005-10-07

申请人： Wei Hu ,  Yunrui Li ,  Juan R. Loaiza

发明人： Wei Hu ,  Yunrui Li ,  Juan R. Loaiza

IPC分类号： G06F15/173 ,  G06F11/00

CPC分类号： H04L67/1097 ,  H04L41/0681 ,  H04L41/0893 ,  H04L43/00 ,  H04L43/0817

摘要： Detecting an operating condition in a multi-server data storage system. One or more of the data storage servers are designated with a status within the cluster. The designating may be performed by at least one of the data storage servers. The status is encoded into log data having records that report changes to the primary data storage performed by the cluster of data storage servers. The log data is transmitted from the primary data storage to the standby data storage system. The standby data storage system determines an operating condition of the cluster of data storage servers based, at least in part, on the status.

摘要翻译： 检测多服务器数据存储系统中的运行状况。 一个或多个数据存储服务器被指定为集群内的状态。 指定可以由数据存储服务器中的至少一个来执行。 该状态被编码成具有报告对由数据存储服务器集群执行的主数据存储的更改的记录的日志数据。 日志数据从主数据存储器传送到备用数据存储系统。 备用数据存储系统至少部分地基于状态来确定数据存储服务器的集群的操作条件。

公开(公告)号：US08438169B2

公开(公告)日：2013-05-07

申请号：US10982135

申请日：2004-11-05

申请人： Wei Hu ,  Yunrui Li ,  Vinay Srihari ,  Ramana Yemeni

发明人： Wei Hu ,  Yunrui Li ,  Vinay Srihari ,  Ramana Yemeni

IPC分类号： G06F17/30

摘要： Each of a plurality of Worker processes are allowed to perform any and all of the following tasks involving logged work items: (1) reading a subset of the work items from a log; (2) sequentially ordering work items for corresponding data objects; (3) applying a sequentially ordered set of work items to a corresponding data object; and (4) transmitting a subset of work items to a Worker process running on another database server in a cluster, if necessary. These tasks can be performed concurrently, at will, and as available, by the Worker processes. An improved checkpointing technique eliminates the need for the Worker processes to get to a synchronization point and stop. Instead, a Coordinator process examines the current state of progress of the Worker processes and computes a past point in the sequence of work items at which all work items before that point have been completely processed, and records this point as the checkpoint.

公开(公告)号：US20070083641A1

公开(公告)日：2007-04-12

申请号：US11245868

申请日：2005-10-07

申请人： Wei Hu ,  Yunrui Li ,  Juan Loaiza

发明人： Wei Hu ,  Yunrui Li ,  Juan Loaiza

IPC分类号： G06F15/173

CPC分类号： H04L67/1097 ,  H04L41/0681 ,  H04L41/0893 ,  H04L43/00 ,  H04L43/0817

摘要： Detecting an operating condition in a multi-server data storage system. One or more of the data storage servers are designated with a status within the cluster. The designating may be performed by at least one of the data storage servers. The status is encoded into log data having records that report changes to the primary data storage performed by the cluster of data storage servers. The log data is transmitted from the primary data storage to the standby data storage system. The standby data storage system determines an operating condition of the cluster of data storage servers based, at least in part, on the status.

摘要翻译： 检测多服务器数据存储系统中的运行状况。 一个或多个数据存储服务器被指定为集群内的状态。 指定可以由数据存储服务器中的至少一个来执行。 该状态被编码成具有报告对由数据存储服务器集群执行的主数据存储的更改的记录的日志数据。 日志数据从主数据存储器传送到备用数据存储系统。 备用数据存储系统至少部分地基于状态来确定数据存储服务器的集群的操作条件。