[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: [lvm-devel] [PATCH v3 01/18] fsadm: Add new commands create, list, add and remove



On Wed, 5 Oct 2011, Alasdair G Kergon wrote:

> On Wed, Oct 05, 2011 at 01:21:17PM +0200, Lukas Czerner wrote:
> > On Wed, 5 Oct 2011, Alasdair G Kergon wrote:
> > >   --force is meant to mean 'permit operations that might not always be wise but
> > > sometimes do need to be done - assume the user knows what they are doing'.
> > > Normally a warning of what will happen is issued and confirmation is required,
> > > but --yes can be used to avoid that confirmation.
>  
> > Do you mean --force in the last sentence ?
>  
> No, I meant --yes.
> 
> It works like this:
> 
>   If you use --yes, you will not be asked any questions.
>   Where the code would otherwise have issued a question, it proceeds as if you
> answered 'yes, do it'.
> 
> If you use --force, it can open up *extra* things the code allows to happen.
> Without it, the code would just give an error 'I'm not going to let you do that'.
> 
> 
> So, the number of --force options supplied controls *what* the code lets you do.
> 
> --yes just controls whether or not you get interactive prompts advising you
> what will happen and giving you chances to bail out.

Ok, now it is clear to me. However --yes option does not exist in some
lvm tolls (e.g. lvremove, vgremove)

> 
> > The first misunderstanding was the use of -f in lvm toos. Because I can
> > not see -y|--yes in the documentation I simply used force for not asking
> > questions about active volumes to be removed. This can be fixed, if '-y'
> > really works.
>  
> I can see things missing from the man pages...
> 
> > The other is, that using -y option to answer yes to umount mounted file
> > system to be checked is ok, but then running fsck with that '-y' option
> > does not make sense and it is also dangerous. But Zdenek is convinced
> > that it has to stay there (I say it is a bug).
> 
> Well what we could do there is:
>   --yes  only:  we do NOT supply -y to fsck.  (Ideally we'd like to choose
> non-interactive options though i.e. if it hits something that would
> normally require an interactive response, then instead of prompting,
> fail the operation.  Perhaps -y should map to 'fsck -n' OR 'fsck -p'?
> It should be non-interactive AND safe.)

Non interactive would be fsck -p which actually will try to fix the file
system, not answer yes to all questions, however I am not sure if we
want it to map it to --yes option.

> 
>   --force --yes: This could perhaps map to -y, if there's a case
> for that.  Or perhaps '--force --force --yes'

The problem here is, that --yes is also used to umount the file system
prior to the fsck. I would be in favor of just not supporting fsck -y in
fsadm, but having fsck -p for non interactive use.

> 
> > lvresize as well, where you have to use "--nofsck" (which is broken btw).
>  
> The reason for --nofsck is if you *know* the filesystem is ok because you
> already ran a check (for example).  But if the fs is indeed ok, won't a basic
> fsck be fast anyway, (just checking fs state in superblock rather than
> performing a full check) such that it doesn't matter if it gets run again?

We can easily get the information whether the file system has been
checked since the last umount and skip the check automatically, since
resize2fs will not require it. However basic check is not very useful,
since as you said it'll only check the state of the super block, which
is not enough and resize2fs will refuse to run in the case that the file
system has not received full check from the last mount.

So as you can see, --nofsck is actually not needed for the case you've
described and I would prefer not allowing user to override that
behaviour and always do fsck prior to the resize2fs, but I can not
insist on it - it just can not be done with --force, but rather with
--nofsck.

> 
> Alasdair

-Lukas


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]