[comp.unix.admin] Do you run Unix without disk quotas?

timcc@csv.viccol.edu.au (02/16/91)

I am soliciting opinions from administrators of Unix systems with a large
number of users (undergraduate students to be specific), but no disk quota
mechanism.

We are currently evaluating a tender that includes Unix systems with and
without disk quotas.  I am interested in answers to the following
questions:

   1.	Have you implemented any other method of controlling disk usage?
	If so, what does it entail?

   1a.	How difficult was the implentation of this alternative?

   1b.	Is it effective?  Is it better or worse for the administrator
	and for the user?

   2.	If you have not implemented an alternative, how much more disk
	space do you think you use (if any)?

   3.	How much more time does the system administrator spend controlling
	disk space usage, either with an alternative method of control, or
	"by hand".

   4.	How would you rate the presence of a disk quota system in
	importance, compared to other system features (for example, dynamic
	disk bad-block re-mapping, an extended access-control mechanism,
	or adherence to contemporary Unix and Open Systems standards).

I would like replies by e-mail.  I will post a precis of what I get to this
newsgroup.  I will be happy to forward a full copy of any answers I receive
to anyone who asks.

--
Tim Cook        Systems Administrator, Victoria College Computer Services
Internet:       timcc@viccol.edu.au
UUCP:           ...!uunet!viccol.edu.au!timcc

schales@photon.tamu.edu (Douglas Lee Schales) (02/22/91)

In article <28961@cs.yale.edu> anselmo-ed@CS.YALE.EDU (Ed Anselmo) writes:
   >>>>> On 15 Feb 91 17:00:46 GMT, timcc@csv.viccol.edu.au said:

   Tim> I am soliciting opinions from administrators of Unix systems with
   Tim> a large number of users (undergraduate students to be specific),
   Tim> but no disk quota mechanism.

   We don't run quota's on any of our machines here.

   Willful ignorance; Public humiliation; daily posting of the stats from
   "df" and a list of the "heavy disk users" (in our case accounts using
   > 5MB) to a local newsgroup.

   When the CS majors' partition fills up, it's their fault.  We'll post
   a quick message to a local newsgroup.  Since none of the majors can
   get any work done, typically space gets freed up quickly.

   They always seem to be able to come up with an extra 15-20 MB on
   demand when the partition fills.  Amazing.

   Non-majors taking CS classes get accounts on another partition.  Still
   no disk quotas, but instructors and TA's typically keep watch on the
   disk usage of their class accounts.


I'll agree to most of that, as I ran a site in exactly this manner.  It *is*
quite amazing how quickly users can clear 15-20MB's when they need it.  The
problem with this is that *1* user can create havoc for all of the other
users on this partition.  It is hardly fair that these users aren't able to
work because one person has filled up the partition.  Even though everyone
may know who is causing the problem, this won't help them with their project
which is due the next day.

We use quotas here.  The biggest headache is keeping track of a person's quota
when they are moved to a different partition... zap the old one quota, create
a quota on the new partition.

Doug.
-----
Douglas Lee Schales
schales@cs.tamu.edu

john@mintaka.mlb.semi.harris.com (John M. Blasik) (02/22/91)

In article <SCHALES.91Feb21111108@photon.tamu.edu> schales@photon.tamu.edu (Douglas Lee Schales) writes:
>In article <28961@cs.yale.edu> anselmo-ed@CS.YALE.EDU (Ed Anselmo) writes:
>   >>>>> On 15 Feb 91 17:00:46 GMT, timcc@csv.viccol.edu.au said:
>
>
>   We don't run quota's on any of our machines here.

Same here at a .com site.
    [ more of the same stuff we do delete...]

>
>I'll agree to most of that, as I ran a site in exactly this manner.  It *is*
>quite amazing how quickly users can clear 15-20MB's when they need it.  The

Not as amazing as how our users can clear 150-200MB!

>problem with this is that *1* user can create havoc for all of the other
>users on this partition.  It is hardly fair that these users aren't able to
>work because one person has filled up the partition.  Even though everyone
>may know who is causing the problem, this won't help them with their project
>which is due the next day.

Sounds like you need some BIGGER, NASTIER, MEANER users with stuff due.
'Round here we have the biggest, nastiest, meanest hoss
(goes by the name "BIG MAC") as ~enforcer.
	% groups mac
	mac : users W

needless to say, most partitions contain directories writeable by group-W :-)

-- john

pavlov@canisius.UUCP (Greg Pavlov) (02/25/91)

In article <1991Feb15.120048.6591@csv.viccol.edu.au>, timcc@csv.viccol.edu.au writes:
> We are currently evaluating a tender that includes Unix systems with and
> without disk quotas.  I am interested in answers to the following
> questions:
> 
>    1.	Have you implemented any other method of controlling disk usage?
> 	If so, what does it entail?
> 
>    1a. How difficult was the implentation of this alternative?
> 
>    1b. Is it effective?  Is it better or worse for the administrator
>        and for the user?
> 
>    2. If you have not implemented an alternative, how much more disk
> 	space do you think you use (if any)?
> 
   We have a reasonably "large" number of users (350-400 at present), many
   of them occasional, but no students.

   Our "philosophy" re disk usage is that, in general, people will waste space
   if allowed unrestricted use but at the same time, using a computer is suffi-
   ciently traumatic for many that another source of hassles ("disk quota
   exceeded....") should be avoided if possible.

   SO: our "alternative" is as follows:

    1. use third-party disks (HP in our case):  purchase cost apx. $3.20/MB .
       These come with 5-year warranties,  so our "maintenance cost" is a
       spare disk to throw into the fray if an on-line unit fails;

    2. use high-volume backup systems.  In our case, Exabyte 2.2 GB per $5.50
       SONY cartridge;

    3. run a program occasionally to check for gross misusage, run another to
       delete certain classes of transient files (logs, "core"s, for instance).

  If things got bad enough we might consider public flagellation (e.g., "the 
  following individuals are anti-social and...").  But this hasn't been nec-
  essary and besides, I've noticed that this sometimes backfires: some people
  see this as a badge of honor....


  pavlov@stewart.fstrf.org

afoiani@nmsu.edu (Anthony "Tkil" Foiani) (02/28/91)

One more small voice...

As an undergrad with a larger-than-normal disk quota, I rarely have
any trouble storing things.  The main culprits to my disk space are
when I want a particular program [say, gnus 3.13, or perl, or ircII]
and the administraters never have the time to install it.

I'd install them myself, but undergrads aren't allowed root access.
More's the pity.

Cheers,
Tony
--
--
Tony Foiani  a.k.a. Tkil  (afoiani@nmsu.edu) or (mcsajf@nmsuvm1.bitnet)
Supporting:  Unix / DOS / VMS / Macintosh / "What's this?"
 "But, although we make records for a hard world, we make records for
  a hard world that you can get by in, and that you can have fun in,
  that you can get off on." - Andrew Eldritch of The Sisters Of Mercy