Le 06/08/17 à 12:55, Petter Reinholdtsen a écrit :
[Laurent Bigonville]
The discards commands will also be issued when shrinking or moving a LV
to an other PV, if something is going wrong during these operations, the
data will be lost.
So it's not only when explicitly removing an LV.
Aha. I d
[Laurent Bigonville]
> The discards commands will also be issued when shrinking or moving a LV
> to an other PV, if something is going wrong during these operations, the
> data will be lost.
>
> So it's not only when explicitly removing an LV.
Aha. I did not have the imagination required to con
Le 06/08/17 à 10:50, Bastian Blank a écrit :
On Sun, Aug 06, 2017 at 09:16:55AM +0200, Petter Reinholdtsen wrote:
A while back, I created a deb named ssd-setup to automatically set up Debian for
a SSD disk. Until I understand more about the LVM issue, I automatically
enable issue_discards=1 in
[Bastian Blank]
> In general it is a bad idea.
I belive this is not clear enough from the documentation and the
instructions on http://wiki.debian.org/SSDOptimization >. But I
am unsure how to make it more clear. :/
I must admit that https://bugs.debian.org/717313#36 > describe my
expectations
On Sun, Aug 06, 2017 at 09:16:55AM +0200, Petter Reinholdtsen wrote:
> A while back, I created a deb named ssd-setup to automatically set up Debian
> for
> a SSD disk. Until I understand more about the LVM issue, I automatically
> enable issue_discards=1 in
> https://anonscm.debian.org/cgit/colla
[Laurent Bigonville]
> What I understand from this bug and Zdenek (who is apparently LVM2 upstream)
> message is that by default the discard requests are already passed to the
> underlying device during regular operation, this was definitely not really
> clear in my head, maybe this should be expla
On Fri, 19 Jul 2013 10:43:11 +0200 Petter Reinholdtsen
wrote:
>
> Hi.
Hello,
> According to several recipes on how to set up Linux on a SSD disk, LVM
> need to be reconfigured to issue discard (aka TRIM) instructions to the
> underlying storage device. This is done by modifying /etc/lvm/lvm.co
Hi,
> Using 'issue_discards=1' makes any lvremove operation 'irreversible' on
> SSD - so using 'vgcfgretore' cannot then fix the mistake of admin when
> he removes LV and he would like to get it back - since restoring TRIMED
> space is just pointless.
>
> So 'default' being 0 is carefully cho
On Sun, 20 Dec 2015 12:50:55 -0800 Matt Taggart wrote:
I agree that "issue_discards = 1" should become the default in lvm on
Debian.
There is a case where you might not want TRIM/discards, there are security
implications to enabling it with dm-crypt. In the commit message which
added this supp
I agree that "issue_discards = 1" should become the default in lvm on
Debian.
There is a case where you might not want TRIM/discards, there are security
implications to enabling it with dm-crypt. In the commit message which
added this support, the author states:
Note that discard will be ne
[Martin Pitt 2013-12-13]
> The documentation says that the option has no effect if the kernel or
> the drive don't support it, so I don't see why we shouldn't just
> enable this by default?
Right. Then perhaps the Debian maintainers of lvm2 can change the
default, to make Debian better for us wi
tag 717313 patch
thanks
Petter Reinholdtsen [2013-07-19 10:43 +0200]:
> According to several recipes on how to set up Linux on a SSD disk, LVM
> need to be reconfigured to issue discard (aka TRIM) instructions to the
> underlying storage device. This is done by modifying /etc/lvm/lvm.conf
> and s
Package: lvm2
Version: 2.02.95-7
Severity: wishlist
Hi.
According to several recipes on how to set up Linux on a SSD disk, LVM
need to be reconfigured to issue discard (aka TRIM) instructions to the
underlying storage device. This is done by modifying /etc/lvm/lvm.conf
and setting issue_discard
13 matches
Mail list logo