[LU-15431] read may return stale data from replicated file

• Type: Bug
• Resolution: Fixed
• Priority: Critical
• Fix Version/s: Lustre 2.16.0
• Labels: flr-improvement
• Severity: 3
• Epic Link: FLR tech debt review
• Assignee: Alex Zhuravlev
• Reporter: Alex Zhuravlev

Description

FLR doesn't hold layout lock (LL) during IO (OST_READ) operations, but verify layout at the beginning of IO and at the end of IO.
the problem is that pages get Uptodate flag from brw_interpret context while the layout is checked for changes later. the pages are still marked uptodate till that point.
three processes:

• 1st modifying the file
• 2nd initiated OST_READ and going to repeat OST_READ due to layout change
• 3rd which finds result of OST_READ (stale data) in ll_do_fast_read()

Issue Links

• is duplicated by: LU-15434 evicted client can return stale data from a mirrored file
• is related to: LU-11695 disabling the xattr cache on client for LSOM