[mdadm PATCH 0/2] two 3.1.3 regression fixes (incremental assembly)

[mdadm PATCH 0/2] two 3.1.3 regression fixes (incremental assembly)

[Dan Williams]

Hi Neil,

The mdadm -I --no-degraded option needs to be acceptable for some time
to come, do you have a policy for when deprecated options can be scheduled for
removal?  The error code for -I shall be zero in the 'not enough' case to
restore the correct and established pre-3.1.3 behavior.

Also available via git:

   git://github.com/djbw/mdadm.git master

---

Dan Williams (2):
      Incremental: return success in 'container not enough' case
      Incremental: accept '--no-degraded' as a deprecated option

 Incremental.c |    2 +-
 mdadm.c       |    2 ++
 2 files changed, 3 insertions(+), 1 deletions(-)

[mdadm PATCH 1/2] Incremental: return success in 'container not enough' case

[Dan Williams]

Commit 97b4d0e9 "Incremental: honor an 'enough' flag from external
handlers" introduced a regression in that it changed the error return
code for successful invocations.

Signed-off-by: Dan Williams <dan.j.williams@intel.com>
---
 Incremental.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/Incremental.c b/Incremental.c
index a8a072d..e4b6196 100644
--- a/Incremental.c
+++ b/Incremental.c
@@ -264,7 +264,7 @@ int Incremental(char *devname, int verbose, int runstop,
 		else {
 			if (verbose)
 				fprintf(stderr, Name ": not enough devices to start the container\n");
-			return 1;
+			return 0;
 		}
 
 		/* This is a pre-built container array, so we do something

[mdadm PATCH 2/2] Incremental: accept '--no-degraded' as a deprecated option

[Dan Williams]

Commit 3288b419 (Revert "Incremental: honor --no-degraded to delay assembly")
killed the --no-degraded flag since commit 97b4d0e9 (Incremental: honor
an 'enough' flag from external handlers) made this the default behavior
of -I, and brought -I usage for external/container formats in line with
native metadata.  However, this breaks existing usages of '-I
--no-degraded', so allow it as a deprecated option.

Starting a degraded container, like the native metadata case, requires -R.

Signed-off-by: Dan Williams <dan.j.williams@intel.com>
---
 mdadm.c |    2 ++
 1 files changed, 2 insertions(+), 0 deletions(-)

diff --git a/mdadm.c b/mdadm.c
index 20a5638..08e8ea4 100644
--- a/mdadm.c
+++ b/mdadm.c
@@ -691,6 +691,8 @@ int main(int argc, char *argv[])
 		"     'summaries', 'homehost', 'byteorder', 'devicesize'.\n");
 			exit(outf == stdout ? 0 : 2);
 
+		case O(INCREMENTAL,NoDegraded):
+			fprintf(stderr, Name ": --no-degraded is deprecated in Incremental mode\n");
 		case O(ASSEMBLE,NoDegraded): /* --no-degraded */
 			runstop = -1; /* --stop isn't allowed for --assemble,
 				       * so we overload slightly */

Re: [mdadm PATCH 0/2] two 3.1.3 regression fixes (incremental assembly)

[Dan Williams]

On 8/10/2010 9:46 AM, Williams, Dan J wrote:
> Hi Neil,
>
> The mdadm -I --no-degraded option needs to be acceptable for some time
> to come, do you have a policy for when deprecated options can be scheduled for
> removal?  The error code for -I shall be zero in the 'not enough' case to
> restore the correct and established pre-3.1.3 behavior.
>
> Also available via git:
>
>     git://github.com/djbw/mdadm.git master
>
> ---
>
> Dan Williams (2):
>        Incremental: return success in 'container not enough' case
>        Incremental: accept '--no-degraded' as a deprecated option

Note, both of these should have been tagged:

Reported-by: Ignacy Kasperowicz <ignacy.kasperowicz@intel.com>

If you take the patches directly.

Thanks,
Dan

Re: [mdadm PATCH 0/2] two 3.1.3 regression fixes (incremental assembly)

[Neil Brown]

On Tue, 10 Aug 2010 16:40:46 -0700
Dan Williams <dan.j.williams@intel.com> wrote:

> On 8/10/2010 9:46 AM, Williams, Dan J wrote:
> > Hi Neil,
> >
> > The mdadm -I --no-degraded option needs to be acceptable for some time
> > to come, do you have a policy for when deprecated options can be scheduled for
> > removal?  The error code for -I shall be zero in the 'not enough' case to
> > restore the correct and established pre-3.1.3 behavior.
> >
> > Also available via git:
> >
> >     git://github.com/djbw/mdadm.git master
> >
> > ---
> >
> > Dan Williams (2):
> >        Incremental: return success in 'container not enough' case
> >        Incremental: accept '--no-degraded' as a deprecated option
> 
> Note, both of these should have been tagged:
> 
> Reported-by: Ignacy Kasperowicz <ignacy.kasperowicz@intel.com>
> 
> If you take the patches directly.

Thanks,
I've taken both and added that line.

I'm beginning to think we should treat 3.1.3 like an -rc and do a 3.1.4 soon
with all these little fix-ups...

NeilBrown

Re: [mdadm PATCH 0/2] two 3.1.3 regression fixes (incremental assembly)

[Neil Brown]

On Tue, 10 Aug 2010 09:46:38 -0700
Dan Williams <dan.j.williams@intel.com> wrote:

> Hi Neil,
> 
> The mdadm -I --no-degraded option needs to be acceptable for some time
> to come, do you have a policy for when deprecated options can be scheduled for
> removal?  The error code for -I shall be zero in the 'not enough' case to
> restore the correct and established pre-3.1.3 behavior.

No, I don't have a policy.  Given that this was introduced fairly recently I
don't feel a need to keep it around for very long.  I'd probably be happy to
see in go in 12 months, or with mdadm-v4.0, whichever is first.

Thanks,
NeilBrown