[PATCH] xfstests: add freeze and dangerous groups

[PATCH] xfstests: add freeze and dangerous groups

[Eric Sandeen]

Add 2 new test groups:

freeze: tests which test filesystem freeze
dangerous: tests which may hang or oops

The 2nd may be useful for automated testing to do i.e.

./check -g auto -x dangerous
./check -g auto,dangerous

to try to get fuller coverage before running into tests
which may panic or hang the box and stop the test cycle.

I doubt I have all the potential dangerous tests, but
they can be added later when found.

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

diff --git a/group b/group
index 01f406d..90b7a84 100644
--- a/group
+++ b/group
@@ -1,4 +1,3 @@
-#
 # QA groups control file
 # Defines test groups and nominal group owners
 # - do not start group names with a digit
@@ -104,6 +103,12 @@ atime
 # Test preallocation calls
 prealloc
 
+# Test filesystem freeze
+freeze
+
+# Tests which may oops or hang
+dangerous
+
 # Old tests that we won't spend any effort trying to run and make work
 # on current systems
 deprecated
@@ -178,7 +183,7 @@ deprecated
 065 dump auto
 066 dump ioctl auto quick
 067 acl attr auto quick
-068 other auto
+068 other auto freeze dangerous
 069 rw udf auto quick
 070 attr udf auto quick
 071 rw
@@ -229,7 +234,7 @@ deprecated
 116 quota auto quick
 117 attr auto quick
 118 quota auto quick
-119 log v2log auto
+119 log v2log auto freeze dangerous
 120 other auto quick
 121 log auto quick
 122 other auto quick
@@ -393,4 +398,4 @@ deprecated
 277 auto ioctl quick metadata
 278 repair auto
 279 auto mkfs
-280 auto quota
+280 auto quota freeze dangerous

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

Re: [PATCH] xfstests: add freeze and dangerous groups

[Dave Chinner]

On Fri, Mar 02, 2012 at 10:56:17AM -0600, Eric Sandeen wrote:
> Add 2 new test groups:
> 
> freeze: tests which test filesystem freeze

That's fine.

> dangerous: tests which may hang or oops

Hmmm.

> The 2nd may be useful for automated testing to do i.e.
> 
> ./check -g auto -x dangerous
> ./check -g auto,dangerous
> 
> to try to get fuller coverage before running into tests
> which may panic or hang the box and stop the test cycle.
> 
> I doubt I have all the potential dangerous tests, but
> they can be added later when found.
> 
> Signed-off-by: Eric Sandeen <sandeen@redhat.com>

I'm not sure "dangerous" is a black and white status for a test.
What if the test doesn't cause problems for upstream, but causes
problems for older vendor kernels? Does that make it dangerous? e.g.
test 104 will hang a RHEL5.x kernel, but is perfectly safe on a
RHEL6.x kernel - does that make it dangerous? It seems that many of
the recent tests for specific regressions fall into this sort of
category. Indeed, how do we answer the question "when does a test no
longer be considered dangerous" or "what test is considered
dangerous for this kernel/platform"?

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

Re: [PATCH] xfstests: add freeze and dangerous groups

[Eric Sandeen]

On 3/4/12 9:08 PM, Dave Chinner wrote:
> On Fri, Mar 02, 2012 at 10:56:17AM -0600, Eric Sandeen wrote:
>> Add 2 new test groups:
>>
>> freeze: tests which test filesystem freeze
> 
> That's fine.
> 
>> dangerous: tests which may hang or oops
> 
> Hmmm.
> 
>> The 2nd may be useful for automated testing to do i.e.
>>
>> ./check -g auto -x dangerous
>> ./check -g auto,dangerous
>>
>> to try to get fuller coverage before running into tests
>> which may panic or hang the box and stop the test cycle.
>>
>> I doubt I have all the potential dangerous tests, but
>> they can be added later when found.
>>
>> Signed-off-by: Eric Sandeen <sandeen@redhat.com>
> 
> I'm not sure "dangerous" is a black and white status for a test.
> What if the test doesn't cause problems for upstream, but causes
> problems for older vendor kernels? Does that make it dangerous? e.g.
> test 104 will hang a RHEL5.x kernel, but is perfectly safe on a
> RHEL6.x kernel - does that make it dangerous? It seems that many of
> the recent tests for specific regressions fall into this sort of
> category. Indeed, how do we answer the question "when does a test no
> longer be considered dangerous" or "what test is considered
> dangerous for this kernel/platform"?

Well, I was thinking that if the original failure mode was a hang or oops, that's "dangerous."

I agree that it's a little nebulous; if you see no value, I'm not hung up on it.

maybe it's a bad choice of words... but the intention was to flag which tests have failure modes which will interrupt further tests.
-Eric

> Cheers,
> 
> Dave.

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

Re: [PATCH] add freeze and dangerous groups

[Christoph Hellwig]

I've applied this one.

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs