Ocfs2 devel - Dec 2008 - [PATCH 1/2] mkfs.ocfs2: Update fs-features list in man page

The list of available options for the fs-features flag in mkfs.ocfs2.8 was
out-dated. Update the list by added information on the 'backup-super',
'inline-data' and 'extended-slotmap' options. Also, update the
reported list
of set features for the --fs-feature-level=default option.

At the same time, we remove the even more out of date list from the
tunefs.ocfs2 man page and provide a pointer to the mkfs.ocfs2 man page. This
way we only have one place to update.

Signed-off-by: Mark Fasheh <mfasheh at suse.com>
---
 mkfs.ocfs2/mkfs.ocfs2.8.in     |   17 ++++++++++++++++-
 tunefs.ocfs2/tunefs.ocfs2.8.in |   24 +-----------------------
 2 files changed, 17 insertions(+), 24 deletions(-)

diff --git a/mkfs.ocfs2/mkfs.ocfs2.8.in b/mkfs.ocfs2/mkfs.ocfs2.8.in
index de43946..8b5dbed 100644
--- a/mkfs.ocfs2/mkfs.ocfs2.8.in
+++ b/mkfs.ocfs2/mkfs.ocfs2.8.in
@@ -108,6 +108,11 @@ Turn specific file system features on or off. A comma
separated list of feature
 The following flags are supported:
 .RS 1.2i
 .TP
+\fBbackup-super\fR
+When this is set, mkfs.ocfs2 will store a copy of the super block in several
known locations so that it can be used for disaster recovery. This feature is
fully compatible with all versions of Ocfs2 and generally shouldn't be
disabled.
+.RE
+.RS 1.2i
+.TP
 \fBlocal\fR
 Create the file system as a local mount, so that it can be mounted without a
cluster stack. Available with the file system bundled with Linux kernels 2.6.20
and later.
 .RE
@@ -126,6 +131,16 @@ Enable support for sparse files. With this, \fIOCFS2\fR can
avoid allocating (an
 \fBunwritten\fR
 Enable unwritten extents support. With this turned on, an application can
request that a range of clusters be pre-allocated within a file. \fIOCFS2\fR
will mark those extents with a special flag so that expensive data zeroing
doesn't have to be performed. Reads and writes to a pre-allocated region act
as reads and writes to a hole, except a write will not fail due to lack of data
allocation. This feature requires \fBsparse\fR file support to be turned on.
Available with the file system bundled with Linux kernels 2.6.23 and later.
 .RE
+.RS 1.2i
+.TP
+\fBinline-data\fR
+Enable inline-data support. If this feature is turned on, Ocfs2 will store
small amounts of file and directory data inside of the inode block. Data is
transparently moved out to an extent when it no longer fits inside the inode
block. In some cases, this can also make a positive impact on cold-cache
directory and file operations.
+.RE
+.RS 1.2i
+.TP
+\fBextended-slotmap\fR
+The slot-map is a hidden file on an Ocfs2 fs which is used to map mounted nodes
to system file resources. The extended slot map allows a larger range of
possible node numbers, which is useful for userspace cluster stacks. This
feature is automatically turned on when needed, thus users have no need to turn
this on manually.
+.RE
 
 .TP
 \fB\-\-fs\-feature\-level=\fR\fR\fIfeature\-level\fR
@@ -138,7 +153,7 @@ Chooses fewer features but ensures that the file system can
be mounted from olde
 .RS 1.2i
 .TP
 \fBdefault\fR
-The default feature set tries to strike a balance between providing new
features and maintaining compatibility with relatively recent versions of
\fIOCFS2\fR. It currently enables \fBsparse\fR.
+The default feature set tries to strike a balance between providing new
features and maintaining compatibility with relatively recent versions of
\fIOCFS2\fR. It currently enables \fBsparse\fR and \fBunwritten\fR.
 .RE
 .RS 1.2i
 .TP
diff --git a/tunefs.ocfs2/tunefs.ocfs2.8.in b/tunefs.ocfs2/tunefs.ocfs2.8.in
index da6c713..96e55d9 100644
--- a/tunefs.ocfs2/tunefs.ocfs2.8.in
+++ b/tunefs.ocfs2/tunefs.ocfs2.8.in
@@ -13,29 +13,7 @@ This tool requires the \fIO2CB\fR cluster to be online.
 .SH "OPTIONS"
 .TP
 \fB\-\-fs\-features=\fR\fR\fI[no]sparse...\fR
-Turn specific file system features on or off. \fItunefs.ocfs2\fR will attempt
to enable or disable the feature list provided. To enable a feature, include it
in the list. To disable a feature, prepend \fBno\fR to the name.
-The following features are supported:
-.RS 1.2i
-.TP
-\fBsparse\fR
-Enable support for sparse files. With this, \fIOCFS2\fR can avoid allocating
(and zeroing) data to fill holes. Available with the file system bundled with
Linux kernels 2.6.22 and later.
-.RE
-.RS 1.2i
-.TP
-\fBunwritten\fR
-Enable unwritten extents support. With this turned on, an application can
request that a range of clusters be pre-allocated within a file. \fIOCFS2\fR
will mark those extents with a special flag so that expensive data zeroing
doesn't have to be performed. Reads and writes to a pre-allocated region act
as reads and writes to a hole, except a write will not fail due to lack of data
allocation. This feature requires \fBsparse\fR file support to be turned on.
Available with the file system bundled with Linux kernels 2.6.23 and later.
-.RE
-
-.TP
-\fB\-J, \-\-journal\-options\fR \fIoptions\fR
-Modify the journal using options specified on the command\-line.
-Journal options are comma separated, and may take an argument using
-the equals ('=') sign. The following options are supported:
-.RS 1.2i
-.TP
-\fBsize\fR=\fIjournal\-size\fR
-Grow the size of journal files for all the slots. Minimum size is 4M.
-.RE
+Turn specific file system features on or off. \fItunefs.ocfs2\fR will attempt
to enable or disable the feature list provided. To enable a feature, include it
in the list. To disable a feature, prepend \fBno\fR to the name. For a list of
feature names, see the man page for \fBmkfs.ocfs2\fR.
 
 .TP
 \fB\-L, \-\-label\fR \fIvolume\-label\fR
-- 
1.5.6

Kernels since 2.6.24 have supported this, I think it's time to make it a
default feature.

Signed-off-by: Mark Fasheh <mfasheh at suse.com>
---
 libocfs2/feature_string.c |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)

diff --git a/libocfs2/feature_string.c b/libocfs2/feature_string.c
index 7b1f93e..e6633da 100644
--- a/libocfs2/feature_string.c
+++ b/libocfs2/feature_string.c
@@ -62,7 +62,8 @@ static struct feature_level_translation
ocfs2_feature_levels_table[] = {
 
 static ocfs2_fs_options feature_level_defaults[] = {
 	{OCFS2_FEATURE_COMPAT_BACKUP_SB | OCFS2_FEATURE_COMPAT_JBD2_SB,
-	 OCFS2_FEATURE_INCOMPAT_SPARSE_ALLOC,
+	 OCFS2_FEATURE_INCOMPAT_SPARSE_ALLOC |
+	 OCFS2_FEATURE_INCOMPAT_INLINE_DATA,
 	 OCFS2_FEATURE_RO_COMPAT_UNWRITTEN},  /* OCFS2_FEATURE_LEVEL_DEFAULT */
 
 	{OCFS2_FEATURE_COMPAT_BACKUP_SB | OCFS2_FEATURE_COMPAT_JBD2_SB,
-- 
1.5.6

Signed-off-by: Tao Ma <tao.ma at oracle.com>

Mark Fasheh wrote:
> The list of available options for the fs-features flag in mkfs.ocfs2.8 was
> out-dated. Update the list by added information on the
'backup-super',
> 'inline-data' and 'extended-slotmap' options. Also, update
the reported list
> of set features for the --fs-feature-level=default option.
> 
> At the same time, we remove the even more out of date list from the
> tunefs.ocfs2 man page and provide a pointer to the mkfs.ocfs2 man page.
This
> way we only have one place to update.
> 
> Signed-off-by: Mark Fasheh <mfasheh at suse.com>
> ---
>  mkfs.ocfs2/mkfs.ocfs2.8.in     |   17 ++++++++++++++++-
>  tunefs.ocfs2/tunefs.ocfs2.8.in |   24 +-----------------------
>  2 files changed, 17 insertions(+), 24 deletions(-)
> 
> diff --git a/mkfs.ocfs2/mkfs.ocfs2.8.in b/mkfs.ocfs2/mkfs.ocfs2.8.in
> index de43946..8b5dbed 100644
> --- a/mkfs.ocfs2/mkfs.ocfs2.8.in
> +++ b/mkfs.ocfs2/mkfs.ocfs2.8.in
> @@ -108,6 +108,11 @@ Turn specific file system features on or off. A comma
separated list of feature
>  The following flags are supported:
>  .RS 1.2i
>  .TP
> +\fBbackup-super\fR
> +When this is set, mkfs.ocfs2 will store a copy of the super block in
several known locations so that it can be used for disaster recovery. This
feature is fully compatible with all versions of Ocfs2 and generally
shouldn't be disabled.
> +.RE
> +.RS 1.2i
> +.TP
>  \fBlocal\fR
>  Create the file system as a local mount, so that it can be mounted without
a cluster stack. Available with the file system bundled with Linux kernels
2.6.20 and later.
>  .RE
> @@ -126,6 +131,16 @@ Enable support for sparse files. With this,
\fIOCFS2\fR can avoid allocating (an
>  \fBunwritten\fR
>  Enable unwritten extents support. With this turned on, an application can
request that a range of clusters be pre-allocated within a file. \fIOCFS2\fR
will mark those extents with a special flag so that expensive data zeroing
doesn't have to be performed. Reads and writes to a pre-allocated region act
as reads and writes to a hole, except a write will not fail due to lack of data
allocation. This feature requires \fBsparse\fR file support to be turned on.
Available with the file system bundled with Linux kernels 2.6.23 and later.
>  .RE
> +.RS 1.2i
> +.TP
> +\fBinline-data\fR
> +Enable inline-data support. If this feature is turned on, Ocfs2 will store
small amounts of file and directory data inside of the inode block. Data is
transparently moved out to an extent when it no longer fits inside the inode
block. In some cases, this can also make a positive impact on cold-cache
directory and file operations.
> +.RE
> +.RS 1.2i
> +.TP
> +\fBextended-slotmap\fR
> +The slot-map is a hidden file on an Ocfs2 fs which is used to map mounted
nodes to system file resources. The extended slot map allows a larger range of
possible node numbers, which is useful for userspace cluster stacks. This
feature is automatically turned on when needed, thus users have no need to turn
this on manually.
> +.RE
>  
>  .TP
>  \fB\-\-fs\-feature\-level=\fR\fR\fIfeature\-level\fR
> @@ -138,7 +153,7 @@ Chooses fewer features but ensures that the file system
can be mounted from olde
>  .RS 1.2i
>  .TP
>  \fBdefault\fR
> -The default feature set tries to strike a balance between providing new
features and maintaining compatibility with relatively recent versions of
\fIOCFS2\fR. It currently enables \fBsparse\fR.
> +The default feature set tries to strike a balance between providing new
features and maintaining compatibility with relatively recent versions of
\fIOCFS2\fR. It currently enables \fBsparse\fR and \fBunwritten\fR.
>  .RE
>  .RS 1.2i
>  .TP
> diff --git a/tunefs.ocfs2/tunefs.ocfs2.8.in
b/tunefs.ocfs2/tunefs.ocfs2.8.in
> index da6c713..96e55d9 100644
> --- a/tunefs.ocfs2/tunefs.ocfs2.8.in
> +++ b/tunefs.ocfs2/tunefs.ocfs2.8.in
> @@ -13,29 +13,7 @@ This tool requires the \fIO2CB\fR cluster to be online.
>  .SH "OPTIONS"
>  .TP
>  \fB\-\-fs\-features=\fR\fR\fI[no]sparse...\fR
> -Turn specific file system features on or off. \fItunefs.ocfs2\fR will
attempt to enable or disable the feature list provided. To enable a feature,
include it in the list. To disable a feature, prepend \fBno\fR to the name.
> -The following features are supported:
> -.RS 1.2i
> -.TP
> -\fBsparse\fR
> -Enable support for sparse files. With this, \fIOCFS2\fR can avoid
allocating (and zeroing) data to fill holes. Available with the file system
bundled with Linux kernels 2.6.22 and later.
> -.RE
> -.RS 1.2i
> -.TP
> -\fBunwritten\fR
> -Enable unwritten extents support. With this turned on, an application can
request that a range of clusters be pre-allocated within a file. \fIOCFS2\fR
will mark those extents with a special flag so that expensive data zeroing
doesn't have to be performed. Reads and writes to a pre-allocated region act
as reads and writes to a hole, except a write will not fail due to lack of data
allocation. This feature requires \fBsparse\fR file support to be turned on.
Available with the file system bundled with Linux kernels 2.6.23 and later.
> -.RE
> -
> -.TP
> -\fB\-J, \-\-journal\-options\fR \fIoptions\fR
> -Modify the journal using options specified on the command\-line.
> -Journal options are comma separated, and may take an argument using
> -the equals ('=') sign. The following options are supported:
> -.RS 1.2i
> -.TP
> -\fBsize\fR=\fIjournal\-size\fR
> -Grow the size of journal files for all the slots. Minimum size is 4M.
> -.RE
> +Turn specific file system features on or off. \fItunefs.ocfs2\fR will
attempt to enable or disable the feature list provided. To enable a feature,
include it in the list. To disable a feature, prepend \fBno\fR to the name. For
a list of feature names, see the man page for \fBmkfs.ocfs2\fR.
>  
>  .TP
>  \fB\-L, \-\-label\fR \fIvolume\-label\fR

Comments inlined.

Mark Fasheh wrote:
> The list of available options for the fs-features flag in mkfs.ocfs2.8 was
> out-dated. Update the list by added information on the
'backup-super',
> 'inline-data' and 'extended-slotmap' options. Also, update
the reported list
> of set features for the --fs-feature-level=default option.
>
> At the same time, we remove the even more out of date list from the
> tunefs.ocfs2 man page and provide a pointer to the mkfs.ocfs2 man page.
This
> way we only have one place to update.
>
> Signed-off-by: Mark Fasheh <mfasheh at suse.com>
> ---
>  mkfs.ocfs2/mkfs.ocfs2.8.in     |   17 ++++++++++++++++-
>  tunefs.ocfs2/tunefs.ocfs2.8.in |   24 +-----------------------
>  2 files changed, 17 insertions(+), 24 deletions(-)
>
> diff --git a/mkfs.ocfs2/mkfs.ocfs2.8.in b/mkfs.ocfs2/mkfs.ocfs2.8.in
> index de43946..8b5dbed 100644
> --- a/mkfs.ocfs2/mkfs.ocfs2.8.in
> +++ b/mkfs.ocfs2/mkfs.ocfs2.8.in
> @@ -108,6 +108,11 @@ Turn specific file system features on or off. A comma
separated list of feature
>  The following flags are supported:
>  .RS 1.2i
>  .TP
> +\fBbackup-super\fR
> +When this is set, mkfs.ocfs2 will store a copy of the super block in
several known locations so that it can be used for disaster recovery. This
feature is fully compatible with all versions of Ocfs2 and generally
shouldn't be disabled.
backup-super already exists below local. And then there is 
--no-backup-super.
All with different wording. How about we remove the other two and tweak the
wording as follows:

\fBmkfs.ocfs2\fR, by default, makes upto 6 backup copies of the super block
at offsets 1G, 4G, 16G, 64G, 256G and 1T depending on the size of the 
volume.
This can be useful in disaster recovery. This feature is fully compatible
with all versions of the file system and generally should not be disabled.
> +.RE
> +.RS 1.2i
> +.TP
>  \fBlocal\fR
>  Create the file system as a local mount, so that it can be mounted without
a cluster stack. Available with the file system bundled with Linux kernels
2.6.20 and later.
>  .RE
> @@ -126,6 +131,16 @@ Enable support for sparse files. With this,
\fIOCFS2\fR can avoid allocating (an
>  \fBunwritten\fR
>  Enable unwritten extents support. With this turned on, an application can
request that a range of clusters be pre-allocated within a file. \fIOCFS2\fR
will mark those extents with a special flag so that expensive data zeroing
doesn't have to be performed. Reads and writes to a pre-allocated region act
as reads and writes to a hole, except a write will not fail due to lack of data
allocation. This feature requires \fBsparse\fR file support to be turned on.
Available with the file system bundled with Linux kernels 2.6.23 and later.
>  .RE
> +.RS 1.2i
> +.TP
> +\fBinline-data\fR
> +Enable inline-data support. If this feature is turned on, Ocfs2 will store
small amounts of file and directory data inside of the inode block. Data is
transparently moved out to an extent when it no longer fits inside the inode
block. In some cases, this can also make a positive impact on cold-cache
directory and file operations.
Enable inline-data support. If this feature is turned on, /fIOCFS2/fR will
store small files and directories inside the inode block. Data is
transparently moved out to an extent when it no longer fits inside the
inode block. In some cases, this can also make a positive impact on
cold-cache directory and file operations. Available with the file system
bundled with Linux kernels 2.6.24 and later.

> +.RE
> +.RS 1.2i
> +.TP
> +\fBextended-slotmap\fR
> +The slot-map is a hidden file on an Ocfs2 fs which is used to map mounted
nodes to system file resources. The extended slot map allows a larger range of
possible node numbers, which is useful for userspace cluster stacks. This
feature is automatically turned on when needed, thus users have no need to turn
this on manually.
> +.RE
Change Ocfs2 to \fIOCFS2\fR. Also append:
Available with the file system bundled with Linux kernels 2.6.27 and later.
>  
>  .TP
>  \fB\-\-fs\-feature\-level=\fR\fR\fIfeature\-level\fR
> @@ -138,7 +153,7 @@ Chooses fewer features but ensures that the file system
can be mounted from olde
>  .RS 1.2i
>  .TP
>  \fBdefault\fR
> -The default feature set tries to strike a balance between providing new
features and maintaining compatibility with relatively recent versions of
\fIOCFS2\fR. It currently enables \fBsparse\fR.
> +The default feature set tries to strike a balance between providing new
features and maintaining compatibility with relatively recent versions of
\fIOCFS2\fR. It currently enables \fBsparse\fR and \fBunwritten\fR.
>  .RE
>  .RS 1.2i
>  .TP
> diff --git a/tunefs.ocfs2/tunefs.ocfs2.8.in
b/tunefs.ocfs2/tunefs.ocfs2.8.in
> index da6c713..96e55d9 100644
> --- a/tunefs.ocfs2/tunefs.ocfs2.8.in
> +++ b/tunefs.ocfs2/tunefs.ocfs2.8.in
> @@ -13,29 +13,7 @@ This tool requires the \fIO2CB\fR cluster to be online.
>  .SH "OPTIONS"
>  .TP
>  \fB\-\-fs\-features=\fR\fR\fI[no]sparse...\fR
> -Turn specific file system features on or off. \fItunefs.ocfs2\fR will
attempt to enable or disable the feature list provided. To enable a feature,
include it in the list. To disable a feature, prepend \fBno\fR to the name.
> -The following features are supported:
> -.RS 1.2i
> -.TP
> -\fBsparse\fR
> -Enable support for sparse files. With this, \fIOCFS2\fR can avoid
allocating (and zeroing) data to fill holes. Available with the file system
bundled with Linux kernels 2.6.22 and later.
> -.RE
> -.RS 1.2i
> -.TP
> -\fBunwritten\fR
> -Enable unwritten extents support. With this turned on, an application can
request that a range of clusters be pre-allocated within a file. \fIOCFS2\fR
will mark those extents with a special flag so that expensive data zeroing
doesn't have to be performed. Reads and writes to a pre-allocated region act
as reads and writes to a hole, except a write will not fail due to lack of data
allocation. This feature requires \fBsparse\fR file support to be turned on.
Available with the file system bundled with Linux kernels 2.6.23 and later.
> -.RE
> -
> -.TP
> -\fB\-J, \-\-journal\-options\fR \fIoptions\fR
> -Modify the journal using options specified on the command\-line.
> -Journal options are comma separated, and may take an argument using
> -the equals ('=') sign. The following options are supported:
> -.RS 1.2i
> -.TP
> -\fBsize\fR=\fIjournal\-size\fR
> -Grow the size of journal files for all the slots. Minimum size is 4M.
> -.RE
> +Turn specific file system features on or off. \fItunefs.ocfs2\fR will
attempt to enable or disable the feature list provided. To enable a feature,
include it in the list. To disable a feature, prepend \fBno\fR to the name. For
a list of feature names, see the man page for \fBmkfs.ocfs2\fR.
>  
>  .TP
>  \fB\-L, \-\-label\fR \fIvolume\-label\fR
--journal-option is not a fs-feature per se. We should leave it as is.
Rest is ok.

sob

man page will need tweaking. ;) I can handle that later.

Mark Fasheh wrote:
> Kernels since 2.6.24 have supported this, I think it's time to make it
a
> default feature.
>
> Signed-off-by: Mark Fasheh <mfasheh at suse.com>
> ---
>  libocfs2/feature_string.c |    3 ++-
>  1 files changed, 2 insertions(+), 1 deletions(-)
>
> diff --git a/libocfs2/feature_string.c b/libocfs2/feature_string.c
> index 7b1f93e..e6633da 100644
> --- a/libocfs2/feature_string.c
> +++ b/libocfs2/feature_string.c
> @@ -62,7 +62,8 @@ static struct feature_level_translation
ocfs2_feature_levels_table[] = {
>  
>  static ocfs2_fs_options feature_level_defaults[] = {
>  	{OCFS2_FEATURE_COMPAT_BACKUP_SB | OCFS2_FEATURE_COMPAT_JBD2_SB,
> -	 OCFS2_FEATURE_INCOMPAT_SPARSE_ALLOC,
> +	 OCFS2_FEATURE_INCOMPAT_SPARSE_ALLOC |
> +	 OCFS2_FEATURE_INCOMPAT_INLINE_DATA,
>  	 OCFS2_FEATURE_RO_COMPAT_UNWRITTEN},  /* OCFS2_FEATURE_LEVEL_DEFAULT */
>  
>  	{OCFS2_FEATURE_COMPAT_BACKUP_SB | OCFS2_FEATURE_COMPAT_JBD2_SB,
>