Computer diagnostic system and method
View Patent ↗A method used on a computer system for determining a most likely cause of a system problem associated with the computer system. By performing the method, the computer system constructs a device tree containing information about all relevant computer devices, determines a score for each of the computer devices represented in the device tree based on predetermined criteria, and aggregates and analyzes the score for each of the computer devices to identify a most likely cause of the system problem. The device tree may be created as having hierarchy levels of tree nodes, each node representing one of all relevant computer devices and each parent node representing a computer device controlling devices represented by child nodes. The score of each of the computer device may be assigned based on whether device is busy, whether the device is idled by the external I/O connection, or whether the device is in a hung or undefined state.
1. A method for troubleshooting a system problem of a computer system, comprising:
constructing at least one data structure containing information about all relevant computer devices organized according to a hierarchy;
determining a score for each of the computer devices in the data structure based on predetermined criteria;
aggregating the score of each of the computer devices to include scores of any associated computer devices lower in the hierarchy; and
analyzing the score for each of the computer devices to identify a most likely cause of the system problem.
2. The method according to claim 1 , wherein the data structure is a tree structure and each node in the tree structure capable of representing a computer device.
3. The method according to claim 1 , wherein the predetermined criteria include at least:
device being busy;
device being idled by the external I/O connection; and
device being in a hung or undefined state.
4. The method according to claim 2 , prior to the step of constructing, further comprising:
creating a crash dump to provide the information about all relevant computer devices.
5. The method according to claim 2 , wherein scoring further comprises:
assigning a score for each leaf node of the tree structure based on the predetermined criteria; and
assigning a score for each non-leaf node of the tree structure as a sum of scores of child nodes of the each non-leaf node.
6. The method according to claim 4 , wherein constructing further comprises:
traversing the crash dump to identify all relevant computer devices;
collecting status information about the devices; and
creating the tree structure having hierarchy levels of tree nodes, each node representing one of all relevant computer devices and each parent node representing a computer device controlling devices represented by child nodes,
wherein the collected status information is recorded in corresponding tree nodes.
7. The method according to claim 5 , wherein analyzing further comprises:
comparing the score of each non-leaf node with scores of child nodes of the each non-leaf node;
determining a non-leaf node with a score greater than each child node's score; and
announcing the non-leaf node as a most likely cause of the system problem.
8. The method according to claim 6 , wherein creating the tree structure further comprises:
representing each node with a full path name including both logical nodes and device nodes, each node separated by a slash (“/”) representing a level of the tree structure hierarchy.
9. A computer-readable medium for use on a computer system, for aggregating system device information to determine a most likely cause of a system problem in the computer system, having computer-executable instructions for performing a method comprising:
constructing at least one device tree containing information about all computer devices organized according to a hierarchy and listed in a previously created crash dump;
determining a score for each of the computer devices in the device tree based on predetermined criteria;
aggregating the score of each of the computer devices to include scores of any associated computer devices lower in the hierarchy; and
identifying the most likely cause of the system problem based upon the aggregated score.
10. The computer-readable medium according to claim 9 , wherein constructing further comprises:
collecting the information from the previously created crash dump; and
creating the device tree having hierarchy levels of tree nodes such that each node represents one of the computer devices and each parent node represents an intermediate computer device controlling computer devices represented by child nodes;
wherein the collected information is recorded in corresponding tree nodes.
11. The computer-readable medium according to claim 9 , wherein the predetermined criteria include at least:
device being busy;
device being idled by the external I/O connection; and
device being in a hung or undefined state.
12. The computer-readable medium according to claim 9 , wherein determining further comprises:
assigning a score for each leaf node of the device tree based on the predetermined criteria; and
assigning a score for each non-leaf node of the device tree as a sum of scores of child nodes of the each non-leaf node.
13. The computer-readable medium according to claim 12 , wherein identifying further comprises:
comparing the score of each non-leaf node with scores of child nodes of the each non-leaf node;
determining a non-leaf node with a score greater than each child node's score; and
announcing the non-leaf node as a most likely cause of the system problem.
14. The computer-readable medium according to claim 13 , wherein identifying further comprises:
announcing a leaf node with the highest score as the most likely cause of the system problem if the score of the non-leaf node is not greater than each of child leaf node's score.