Automatic recovery of a failed standby database in a cluster
US9514160B2 · kind B2 · utility
Assignee
Inventors
Key dates
| Filing date | Mar 11, 2013 |
| Grant date | Dec 6, 2016 |
| Priority date | — |
| Expiry date | Nov 14, 2033 |
Classification
- Technology area (CPC G)Physics
- CPC primaryG06F16/275
- WIPO fieldComputer technology
- WIPO sectorElectrical engineering
Abstract
A method, system, and computer program product. The method for non-intrusive redeployment of a standby database facility comprises configuring a database system having a shared lock manager process to synchronize two or more concurrent access instances, then granting lock requests for access to a cache of database blocks. At some moment in time, the shared lock manager process may fail, and a monitor process detects the failure or other stoppage of the shared lock manager process. A new shared lock manager process and other processes are started, at least one of which serves for identifying the database blocks in the cache that have not yet been written to the database. The identified blocks are formed into a recovery set of redo operations. During this time, incoming requests for access to the cache of database blocks are briefly blocked, at least until the recovery set of redo operations has been formed.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.