lfs at aba.eclipse.co.uk
Tue Aug 27 09:33:09 PDT 2013
On 27/08/13 14:12, Bruce Dubbs wrote:
> David Brodie wrote:
>> On 27/08/13 05:29, Bruce Dubbs wrote:
>>> Fernando de Oliveira wrote:
>>>> Em 26-08-2013 23:58, Fernando de Oliveira escreveu:
>>>>> Tests attached.
>>> ## 417 tests 15:24 : 78 OK, 2 warnings, 20 failures, 0 known failures;
>>> 317 skipped
>>> doesn't seem very useful. You start out with Running
>>> normal:api/lvtest.sh ... passed.
>>> 0:02 0:00.159/0:00.699 14 4536/1448
>>> But that's what hangs for me.
>>> I notice later it says:
>>> kernel_at_least 2 6 33 || skip, but I suspect it doesn't take into
>>> account version 3 of the kernel. Also, there are comments about udev.
>>> Finally, I see snapshot support for the kernel seems to be needed.
>>> Overall, the log does not make much sense to me.
>> The errors look similar to the results I got when running the tests
>> under a 3.9.7 kernel, but with LVM built using 3.8.1 headers - perhaps
>> the API has changed?
> What would change that the tests hang/fail, but the commands seem to
> work fine? I suspect the problem to be something about the test harness.
> -- Bruce
Well I've tried creating some lv's from the command line under linux
3.9.7 (lvm 2.02.98 built with 3.8.1 headers), and the results are
consistent with the test suite, i.e. I can create plain vanilla lv's,
but not with --type=raid/snapshot/mirror, which all give kernel target
not found errors.
I can also create a plain lv on a vg consisting of 3 loopback volumes,
and format, mount, copy files.
Ah, just checked /usr/lib/modules/3.9.7/modules.alias - it's grown
considerably since 3.8.1, and has a whole lot of aliases such as:
alias dm-snapshot-merge dm_snapshot
alias dm-snapshot-origin dm_snapshot
alias dm-raid6 dm_raid
alias dm-raid5 dm_raid
and so on. Wonder if this has something to do with it? lvm2 seems to
think that the kernel target is missing, perhaps it has just changed
name, and lvm2 isn't checking the alias file.
More information about the blfs-dev