xfs/015: enlarge the initial fs size
authorEryu Guan <eguan@redhat.com>
Thu, 4 Jan 2018 05:46:26 +0000 (13:46 +0800)
committerEryu Guan <eguan@redhat.com>
Thu, 15 Feb 2018 17:33:16 +0000 (01:33 +0800)
Kernel commit f59cf5c29919 ("xfs: remove "no-allocation"
reservations for file creations") is known to cause earlier ENOSPC
conditions, and xfs/015 is affected in the way that no new file/dir
can be created in a newly created 16M XFS with both reflink and
rmapbt enabled, thus xfs/015 fails due to the missing test dirs.

So enlarge the initial fs size to 32M that we're able to create new
dir/file before growing the filesystem size.

Note that we doubled the fs initial size, we need to double the
required space too.

Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>
tests/xfs/015

index 2c57f7882308fe0454328762e461e105ea5d5343..4939bb92f6fc78db2f4ccedba658e5e425475a19 100755 (executable)
@@ -66,15 +66,15 @@ _supported_os Linux
 
 _require_scratch
 
-# need 64M space, don't make any assumption
+# need 128M space, don't make any assumption
 _scratch_mkfs >/dev/null 2>&1
 _scratch_mount
-_require_fs_space $SCRATCH_MNT 65536
+_require_fs_space $SCRATCH_MNT 131072
 _scratch_unmount
 
 rm -f $seqres.full
 
-_scratch_mkfs_sized $((16 * 1024 * 1024)) > $tmp.mkfs.raw
+_scratch_mkfs_sized $((32 * 1024 * 1024)) > $tmp.mkfs.raw
 cat $tmp.mkfs.raw | _filter_mkfs >$seqres.full 2>$tmp.mkfs
 # get original data blocks number and agcount
 . $tmp.mkfs