Bug#814644: [lvm2] Stupid changes in /etc/lvm/lvm.conf again and again
zdenek.kabelac at gmail.com
Wed Mar 9 22:11:30 UTC 2016
On Sat, 13 Feb 2016 17:52:36 +0100 Philipp Klaus Krause <pkk at spth.de> wrote:
> Package: lvm2
> Version: 2.02.141-2
> Severity: normal
> --- Please enter the report below this line. ---
> When I upgrade the lvm2 package through synaptic, it comes with a
> /etc/lvm/lvm.conf. Since lvm2 changes the file all the time and I have
> set issue_discards=1 in my /etc/lvm/lvm.conf I always get the dialog
> about the changed config file, and get a list of changes. I have seen
> this multiple times recently; basically on each lvm2 upgrade.
> And when I see the list of changes it is just or mostly indentation or
> linebreaks. Sometimes some real change is hidden in there, too. But lvm2
> should stop making pointless changes that just put additional
> maintanance burden on the user.
> Just as I'm typing this I am doing an lvm2 upgrade; most of the changes
> in the file are bullshit like this:
> - # is not defined, a default single-entry list containing '@*' is
> - # assumed.
> + # is not defined, a default single-entry list containing '@*'
> + # is assumed.
lvm2 package now provides internal support for diffing or showing new options
in differnt releases...
See 'man lvmconfig' for large variety of options.
It may easily generate new config file from scratch.
User may just keep his changes in small file and let other things at defaults.
IMHO much bigger issue is in fact your mentioned 'issue_discard' used - this
option is often misunderstood.
It has nothing in common with TRIM on exiting LV.
It only 'TRIM' space in VG AFTER LV is lvremoved - this makes mostly sense
only in case VG is on some 'provisioned' device.
When such option is 'enabled' user cannot recover LV with 'vgcfgrestore' if
he makes a mistake - while with disabled - data would be still there - and
I've already seen some disappointed users (as Ubuntu even set this to 1 on
So if anyone is using this option with hope it helps sending discard to
activated LVs - then this is not the case - such LVs are passing discard
through if the underlaying PV has discard support.
More information about the pkg-lvm-maintainers