[lvm-devel] configuration details [was Re: vgrename omit renaming hidden volumes if vg is inactive -> thinpools will be inconsistent]

Oliver Rath rath at mglug.de
Fri Aug 8 14:30:24 UTC 2014


Hi Peter,

lvm2 was compiled in this way:

 ./configure \
      --with-thin=internal \
      --with-replicators=internal \
      --enable-udev_sync \
      --enable-udev_rules \
      --enable-python-bindings \
      --enable-applib \
      --enable-cmdlib 

make && make install

Regards
Oliver

On 08.08.2014 10:51, Oliver Rath wrote:
> Hi Peter!
>
> Am 08.08.2014 um 08:26 schrieb Peter Rajnoha:
>> On 08/07/2014 09:40 PM, Oliver Rath wrote:
>>> Hi list,
>>>
>>> ive probably found a bug in lvm2 (testet on commit hash
>>> 6320c3b9052004397faf1687198756ac7914fecc):
>>>
>>> if i rename an inactive vg (vgchange -an vg), then the hidden volumes
>>> like /dev/mapper/vg-mypool_tdata ..tmeta ..tpool arent renamed in
>>> /dev/mapper/..
>>>
>> Inactive VG should not have any /dev/ content. If it remained after
>> deactivation, it seems to be a bug in node/symlink handling under /dev.
>>
>> A few questions then:
>>
>>   - what distro is this?
> Ubuntu 14.04.1 Server amd64 + ubuntu-desktop + lvm2-2.02.106 from git +
> thin-provisioning-tools-0.3.2 from git
>
>>   - are you using udev rules from lvm2 git repo too?
> Yes
>>   - what are the lvm.conf values for these settings:
>>       - activation/udev_sync
> yes
>>       - activation/udev_rules
> yes
>>       - activation/verify_udev_operations
> yes
>>   - what udev (or systemd+udev) version is being used
>>   - are all the items found in /dev/mapper symlinks (not direct block dev nodes)?
>>   - are there dangling symlinks found in /dev/mapper for the LVs from deactivated VG?
>>
> udev-204-5ubuntu2
>
> Im using thinpool-devices from a 700MB thinpool-device.
>
> If you need more, tell me!
>
> Tfh
> Oliver
>
> --
> lvm-devel mailing list
> lvm-devel at redhat.com
> https://www.redhat.com/mailman/listinfo/lvm-devel




More information about the lvm-devel mailing list