I use LTP and our own regression set of scripts designed specifically for unionfs: it sets up several branches/layers, some rw/ro, manually creates files/whiteouts and such on the lower branches, performs some ops via unionfs, and then checks to see if what you see via unionfs is what it’s supposed to be. This unionfs regression is here: http://git.fsl.cs.sunysb.edu/?p=unionfs-regression.git;a=summary
Can you try your xfstests w/ two ext4 layers, just to see if there’s any issue w/ tmpfs? I always first run LTP on the native file systems, record what tests passed/failed, and then put unionfs/wrapfs on top, looking for new failed tests. Thanks, Erez. On May 9, 2014, at 10:50 AM, Theodore Tso <[email protected]> wrote: > Great! Out of curiosity, what are you using for your testing? We > (specifically, Vaibhav) did some tests using a modified version of xfstests, > using tmpfs for a rw and a ro layer, and found the following:
_______________________________________________ unionfs mailing list: http://unionfs.filesystems.org/ [email protected] http://www.fsl.cs.sunysb.edu/mailman/listinfo/unionfs
