EnhanceIO(TM) caching driver features [2/3]

[OS Engineering]

Hi,

Here is a description of error handling as stated in a previous email.

When an IO is submitted to a cache and an OS crash or an abrupt power failure occurs before the cahe completes the IO and returns success status  - The end result after a reboot is that some of the blocks contained in the IO range could be persistent. The rest may not be persistent. In no case will garbage data be found in the IO range. This is in-line with the behavior of RAID systems or volume management software packages. Enterprise class filesystems and databases are capable of handling this situation correctly.

When an SSD fails, dirty data is lost. This may cause an entire filesystem or a database to go corrupt. Administrators however are able to limit exposure to dirty data by using EnhanceIO(TM) automatic clean-up tunables and a forced clean-up.

Attached spreadsheet gives an exhaustive list of the error conditions EnhanceIO(TM) driver handles.
-Amit

Attachment: eioerrorhandling.ods
Description: eioerrorhandling.ods