[PATCH] xfstests: make fs for 274 larger

[PATCH] xfstests: make fs for 274 larger

[Josef Bacik]

Btrfs will default to mixed block groups for 1 gigabyte file systems and
smaller, which means data and metadata share the same area.  This makes
generic/274 fail for us because we cannot reserve enough metadata space to do
our writes.  Bumping the scratch fs up to 2 gigabytes allows us to do our normal
metadata/data seperation and allows us to pass this test.  Thanks,

Signed-off-by: Josef Bacik <jbacik@fusionio.com>
---
 tests/generic/274 |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/tests/generic/274 b/tests/generic/274
index da45fab..7c4887f 100755
--- a/tests/generic/274
+++ b/tests/generic/274
@@ -57,7 +57,7 @@ echo "------------------------------"
 rm -f $seqres.full
 
 umount $SCRATCH_DEV 2>/dev/null
-_scratch_mkfs_sized $((1 * 1024 * 1024 * 1024)) >>$seqres.full 2>&1
+_scratch_mkfs_sized $((2 * 1024 * 1024 * 1024)) >>$seqres.full 2>&1
 _scratch_mount
 
 # Create a 4k file and Allocate 4M past EOF on that file
-- 
1.7.7.6

Re: [PATCH] xfstests: make fs for 274 larger

[Eric Sandeen]

On 6/21/13 3:30 PM, Josef Bacik wrote:
> Btrfs will default to mixed block groups for 1 gigabyte file systems and
> smaller, which means data and metadata share the same area.  This makes
> generic/274 fail for us because we cannot reserve enough metadata space to do
> our writes.  Bumping the scratch fs up to 2 gigabytes allows us to do our normal
> metadata/data seperation and allows us to pass this test.  Thanks,

It'd be nice if we could force btrfs into the other mode just for this
test, but it's probably harmless to just make everyone else bigger.
Makes the test take longer though, another gig to write.

Did you test other filesystems w/ the change?

Seems like a reasonable & expedient solution though, so:

Reviewed-by: Eric Sandeen <sandeen@redhat.com>

> Signed-off-by: Josef Bacik <jbacik@fusionio.com>
> ---
>  tests/generic/274 |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/tests/generic/274 b/tests/generic/274
> index da45fab..7c4887f 100755
> --- a/tests/generic/274
> +++ b/tests/generic/274
> @@ -57,7 +57,7 @@ echo "------------------------------"
>  rm -f $seqres.full
>  
>  umount $SCRATCH_DEV 2>/dev/null
> -_scratch_mkfs_sized $((1 * 1024 * 1024 * 1024)) >>$seqres.full 2>&1
> +_scratch_mkfs_sized $((2 * 1024 * 1024 * 1024)) >>$seqres.full 2>&1
>  _scratch_mount
>  
>  # Create a 4k file and Allocate 4M past EOF on that file
> 

Re: [PATCH] xfstests: make fs for 274 larger

[Josef Bacik]

On Fri, Jun 21, 2013 at 04:53:54PM -0500, Eric Sandeen wrote:
> On 6/21/13 3:30 PM, Josef Bacik wrote:
> > Btrfs will default to mixed block groups for 1 gigabyte file systems and
> > smaller, which means data and metadata share the same area.  This makes
> > generic/274 fail for us because we cannot reserve enough metadata space to do
> > our writes.  Bumping the scratch fs up to 2 gigabytes allows us to do our normal
> > metadata/data seperation and allows us to pass this test.  Thanks,
> 
> It'd be nice if we could force btrfs into the other mode just for this
> test, but it's probably harmless to just make everyone else bigger.
> Makes the test take longer though, another gig to write.
> 
> Did you test other filesystems w/ the change?
> 
> Seems like a reasonable & expedient solution though, so:
> 
> Reviewed-by: Eric Sandeen <sandeen@redhat.com>

Yeah I tested xfs and ext4 and they worked out right, it takes a little longer
but its like 10-15 seconds so not too big of a deal.  I had thought of making
this test force btrfs to do non-mixed but I didn't want to wire all that up when
we could just jack the size up.  Thanks,

Josef