Potential undetected data corruption in IBM Storage Scale when compressed snapshot files are read and the highly-available write cache (HAWC) feature is enabled causing unexpected data reads

Description

When a compressed file in a snapshot is read, the decompression of the data requires writing the data to disk and updating the disk addresses asynchronously. During this process, the data from blocks that are smaller than HAWC write cache threshold are not being flushed to the disk; instead, those smaller data blocks are being appended to the recovery log. Because the disk addresses are updated for the block before the data is flushed to disk, the system attempts to read a disk address that has not yet been flushed. As a result, file contents are read incorrectly for files in snapshots when HAWC and compression are used, resulting in undetected data corruption.

Users Affected:
This issue applies to all supported releases and affects customers that use snapshots with file compression and have HAWC enabled on their file system. 

Problem Determination:
The problem can be determined by reading from a compressed file from a snapshot with HAWC enabled on the file system. No external indicators (like log messages) exist. Reading corrupted data from the snapshot and verifying its inaccuracy is the only action to identify the issue.