发明名称 |
FILE CORRUPTION RECOVERY IN CONCURRENT DATA PROTECTION |
摘要 |
An incremental backup system that performs the following (not necessarily in the following order): (i) making a plurality of time-ordered journal entries; (ii) determining that a corruption condition exists; (iii) responsive to a corruption condition, constructing a first incremental mirror data set that reflects a backup data set and all journal entries up to a first corrupted journal entry which is the earliest in time journal entry, of the plurality of journal entries, that is a corrupted journal entry; (iv) responsive to a corruption condition, constructing a second incremental mirror data set that reflects the backup data set and all journal entries up to the first corrupted journal entry; and (v) checking for corruption in the first and second incremental mirror data sets to determine the latest uncorrupted version of the data set. |
申请公布号 |
US2015178171(A1) |
申请公布日期 |
2015.06.25 |
申请号 |
US201314139906 |
申请日期 |
2013.12.24 |
申请人 |
International Business Machines Corporation |
发明人 |
Bish Thomas W.;Boyd Kenneth W.;McBride Gregory E.;Welp Richard A. |
分类号 |
G06F11/14 |
主分类号 |
G06F11/14 |
代理机构 |
|
代理人 |
|
主权项 |
1. A method comprising:
making a plurality of time-ordered journal entries for incrementally backing up a dynamic source file as a backup data set; determining that a corruption condition exists; responsive to detection of the corruption condition, constructing a first incremental mirror data set that reflects the backup data set and all journal entries up to, but not including, a first corrupted journal entry which is the earliest in time journal entry, of the plurality of journal entries, that is a corrupted journal entry; responsive to detection of the corruption condition, constructing a second incremental mirror data set that reflects the backup data set and all journal entries up to, and including, the first corrupted journal entry; and checking for corruption in the first and second incremental mirror data sets to determine that the first incremental mirror data set is a most up to data set which: (i) is uncorrupted, and (ii) can be constructed from the backup data set and the plurality of time-ordered journal entries; wherein: at least the checking for corruption performed by machine logic implemented by a machine. |
地址 |
Armonk NY US |