linux-ext4 - Re: [PATCH 8/9] e2scrub_all: refactor device probe loop

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20190321194901.44eeu6l7mxsd7k4j@work>
Date:   Thu, 21 Mar 2019 20:49:01 +0100
From:   Lukas Czerner <lczerner@...hat.com>
To:     Theodore Ts'o <tytso@....edu>
Cc:     Ext4 Developers List <linux-ext4@...r.kernel.org>,
        darrick.wong@...cle.com
Subject: Re: [PATCH 8/9] e2scrub_all: refactor device probe loop

On Thu, Mar 21, 2019 at 01:48:11PM -0400, Theodore Ts'o wrote:
> On Thu, Mar 21, 2019 at 11:27:42AM +0100, Lukas Czerner wrote:
> > 
> > Snapshot of a thinspanshot is allowed though, so we might want to
> > include those. Not sure if it's wise to do it by default, but regardless
> > it's probably something for a separate change.
> 
> Yeah, it's definitely a separate change.  One potential design
> question is that for a thin volume, you can do both a thin or a think
> snapshot, and in some cases one might succeed while the other will
> fail.  So do we make this choice be a parameter that we set in the
> config file, or do we try to see if there is sufficient spare
> freespace for a thick snapshot (and then do that), or a thin snapshot
> (and then do that) --- and which should use prefer?

That's a good question. I am not sure it makes sense to allow choice
between thin/thick snapshots of thin volume. Just pick one, ideally it
would be thin snapshot of thin volume.

> 
> The other thing I'll note is that in order for us to tell whether
> something is a thin or thick LV, we're going to to need to ask lvs to
> return multiple parameters, so the optimization of using:
> 
> 	for NAME in $(lvs -o lv_path --noheadings -S...) ; do
> 	    ...
> 	done
> 
> will no longer work.  (Or we end up calling lvs a second time, which
> is less efficient.)

Yeah, however calling lvs twice might still be better then calling
lsblk for every lv ? This can be easily benchmarked.

first it would be for a non-thin volumes so

-S lv_active=active,lv_role=public,lv_role!=snapshot,segtype!=thin

and second for thin volumes and thin snapshots

-S lv_active=active,lv_role=public,segtype=thin

Note that by default thin snapshots are not active so they will not be
checked unless user activated them.

> 
> Just curious --- do we know how commonly thin LV's are being used by
> customers of various distros?  I assume enterprise distro users will
> be the most conservative, but how common is the uptake of thin LV's by
> Fedora and OpenSuSE users?

I wish I knew. I think that for workstation users it's the defaults that
matters the most so I would not expect thin LV's to be used much there.
Qubes OS being most notable exception I think.
I'd think that thin would be mostly used for virtualization and storage
appliances. I'll ask around, Fedora should have some stats.

-Lukas

> 
> 						- Ted