Method and apparatus for identifying failure module
AssigneeFujitsu Limited
InventorsKoji Yamaguchi, Hiroaki Sato, Mikio Ito
Life of this patent
Estimating the term…
Abstract
A failure-data register unit receives failure data from a failure detecting module. Based on the failure data, a test-execution instructing unit instructs each module to execute access tests for the module accessed when the failure detecting module detected the failure. When a test-result receiving unit receives results of the access tests, a failure-location identifying unit identifies the failed module based on the results. A server-request control unit suspends server I/O while the access tests are performed.
Claims13
An apparatus for identifying a failure module in a disk controller that includes a plurality of modules connected on data paths, the apparatus being implemented on a computer executing a computer program and comprising:
- a test-instructing unit that instructs, when one of the modules detects a failure, the modules to execute an access test to a memory with test data;
- and
- an identifying unit that identifies a module as a possibly failed module based on results of the access test, wherein
- when more than one module is detected to have failed, the identifying unit identifies a common module that is commonly connected nearest to and upper than the more than one failed module on the data paths as the possibly failed module,
- when only one of the modules is detected to have failed and when a nearest upper module of the only one of the modules detected on the data paths has no other lower module than the only one of the modules detected on the data paths, the identifying unit identifies both of the nearest upper module and the only one of the modules as the possibly failed module, and
- when only one of the modules is detected to have failed and when a nearest upper module of the only one of the modules has other lower module than the only one of the modules on the data paths, the identifying unit identifies the only one of the modules as the most possibly failed module.
A method for an apparatus to identify a failure module in a disk controller, the apparatus being implemented on a computer by executing a computer program, and the disk controller including a plurality of modules connected on data paths, the method comprising:
- instructing, when one of the modules detects a failure, the modules to execute an access test to a memory with test data;
- and
- identifying a module as a possibly failed module where the failure occurred based on results of the access test, wherein
- when more than one module is detected to be failed at the identifying, a common module that is commonly connected nearest to and upper than the more than one module on the data paths is identified as the possibly failed module,
- when only one of the modules is detected to have failed at the identifying, and when a nearest upper module of the only one of the modules on the data paths has no other lower module than the only one of the modules on the data paths, both of the nearest upper module and the only one of the modules are identified as the possibly failed module, and
- when only one of the modules is detected to have failed at the identifying, and when a nearest upper module of the only one of the modules has other lower module than the only one of the modules detected on the data paths, the only one of the modules is identified as the possibly failed module.
A failure detecting device to identify a failed module among interconnecting modules controlling access to a data storage system, comprising:
- a memory storing test data;
- a control section that upon receiving information that a module failure has occurred, instructs each of the interconnecting modules to execute an access test, by performing at least one of reading and writing first data of the test data and second data of the test data alternatively, and identifies a module among the interconnecting modules that failed based on outcome of the access test;
- and
- a communication interface intermediating communication between the interconnecting modules and the control section or the memory, wherein
- the control section identifies a commonly connected module nearest to and upper than more than one failed module on the data paths, as a module possibly to have failed when multiple module failures are detected.
Dependencies are inferred from the claim text.