Re: Command line mounting all partitions of pluggable device
On Fri 18 Jan 2019 at 07:45:56 (-0600), Richard Owlett wrote:
> On 01/15/2019 10:40 AM, David Wright wrote:
> > On Mon 14 Jan 2019 at 19:06:29 (-0600), Richard Owlett wrote:
> > > On 01/14/2019 02:18 PM, David Wright wrote:
> > > > On Mon 14 Jan 2019 at 10:20:51 (-0600), Richard Owlett wrote:
> > > > > On 01/14/2019 09:22 AM, Thomas Schmitt wrote:
> > > > > > rhkramer@gmail.com wrote:
> > > > > > > > Assuming there are entries in fstab for each partition, [...]
> > > > > > > > for i in /dev/sd<device_letter>*; do mount $i; done
> > > > > >
> > > > > > Rchard Owlett wrote:
> > > > > > > There are not.
> > > > > >
> > > > > > In the most general case i would have a where-to-mount directory with
> > > > > > lots of directories for the various partitions (here 10 drives with
> > > > > > 20 partitions each).
> > > > >
> > > > > But thankfully I don't have to deal with the most general case as I am
> > > > > as compulsive about giving all partitions a reasonably unique label as
> > > > > Debian is about assigning UUIDs.
> > > >
> > > > "Reasonably unique" doesn't really cut it.
> > >
>
> NOTE: When I posted this list, I used tabs rather than spaces to
> format the list. I've reformatted the list so it should survive post &
> quotation problems.
Perhaps "FreeAgent" went walkabout from line 6?
> > > For my current 2.2TB, the labels seem acceptably unique:
> > > MainCdrive my-big-dvd fromdell richardofdell debversionrichar
> > > Projects F_drive OldMachine fullstretch debian8.6
> > > recover-common tomboy-testing owlcommon new-net-inst good-fvwm
> > > tst_mysql tst_mariadb dummy target gddrescued_commo
> > > jessie8-6-6 common stretch-2nd 17oct2017 common-bak
> > > backup_homedirs myhome dectest scratch_pad
> > > FreeAgentGoFlexDrive windows GOFLEXPART5 recovered mate-full
> > > post-failure MISC-backups
> > >
> > Sure, but this sort of ad hoc naming "scheme" means that you have to
> > be aware of the entire current population of names every time you
> > create a new one. That presents a danger of collision.
>
> If partition labels were not explicitly assigned on an _ad hoc basis_,
> they would be useless.
One could debate the merits of any particular naming scheme all day,
but your statement that any planned or systematic scheme is "useless"
seems very odd. But I applaud your consistency: even your choice of
punctuation and capitalisation appears to be ad hoc.
> Consider the usefulness of a dictionary arranged by UUID rather than
> on an _ad hoc basis_ ;/
Not very different, if you're talking about conventional dictionaries
(rather than, say, hashes and other technical meanings).
Dictionary headwords are normally arranged in alphabetical order and,
beneath those headwords, each entry is arranged by grammatical category,
like verb transitive, verb intransitive, noun, adjective, etc.
A dictionary arranged by UUID would effectively be random.
But I think what you're revealing here (correct me if I'm wrong) is
that you've chosen to use partition LABELs as "headwords" for the
information contained in each partition.
> Quoting [https://corporatefinanceinstitute.com/resources/knowledge/terms/ad-hoc/]
>
> > The term “ad hoc” is a Latin phrase that literally means “to this” and
> > commonly understood as meaning “for this purpose.” It can also be used
> > to mean “as needed.” It is commonly used in both business and government
> > settings. Ad hoc refers to actions taken to address a specific situation,
> > circumstance, or problem, and not usually intended to address other or
> > ongoing issues.
>
> IOW “for this purpose” == "finding partition with needed content".
That's your choice of *purpose* for the LABEL; it doesn't follow
from the referenced page. Applying that quotation to this thread,
"as needed" refers to the *process* of choosing LABEL names (ad hoc
as opposed to planned). The purpose, the "needed content", is needed
because, one assumes, that's the content you need to lay your hands on
now or at some time in the future. Your purpose is not ad hoc.
Mentioning other methods of finding content would widen this
discussion excessively, but describing the contents of a partition
in 16 characters (your purpose) does not seem to be particularly
sensible.
> If the labels were not "acceptably unique" they would, by definition,
> *NOT* be "ad hoc" [i.e. not suitable for intended purpose].
That doesn't make sense: "ad hoc" can of course mean "suitable
for intended purpose". It doesn't follow that "Not ad hoc" means
"unsuitable for intended purpose".
(Uniqueness doesn't come into the meaning of ad hoc.)
Using the definition at the top of that web page, your collection
of LABELs certainly looks like a collection of "one off" choices
to me, which is why I described your process of choosing them as
ad hoc.
My own method of choosing LABELs is also vulnerable to name collision,
but because every LABEL starts with the device's name (chosen in an
ad hoc set of systematic ways, and physically written on it when
acquired), collision is unlikely to occur.
Cheers,
David.
Reply to: