SUMMARY: Understatement in Quota's File Counts

From: Jochen Bern (bern@kleopatra.Uni-Trier.DE)
Date: Wed Jan 26 1994 - 01:44:46 CST


(Oooops, this Summary is LATE ...)

I asked:

> introducing some new Users, I had to update the Quotas. I ran a repquota -a -v
> to find a good Prototype User. Some of the Output Lines read:
> Block limits File limits
> User [...] used soft hard timeleft
> anna -- [...] 834 0 0
> gergov -- [...] 1246 0 0
> bern -- [...] 983 0 0
> jordan -- [...] 0 0 0
> etc. etc.. I then did some edquota -p's for the new Users and issued a
> quotacheck -a -v which gave me surprising Results:
> /dev/rsd6h: anna fixed: files 834 -> 1614
> /dev/rsd6h: gergov fixed: files 1246 -> 1841
> /dev/rsd6h: bern fixed: files 983 -> 7932
> /dev/rsd6h: jordan fixed: files 0 -> 98
> etc. etc., in Total, 17 of 25 Users had correct Block Counts but wrong File
> Counts. Note that I haven't set File Number Quotas. The most surprising ones
> were Pseudo-Users with few Files like jordan - I'm pretty sure that I have
> seen them with a non-Zero Count the last Time I did a quotacheck. What causes
> these wrong Counts, is it a known Problem?
> Environment: SunOS 4.1.2, OW 3.0, NFS Server a SS2, Clients mainly ELCs,
> Home Dirs exported -root=<explicit List>,access=<same List>,
> mounted rw,intr,bg.

The Answer:

1) Repquota is broken in that it doesn't give you the correct Numbers.
   Quotas of single Users can be gotten correctly by quota -v. Available
   Patches are 100965-01 (quota console msgs), 101039-09 (repquota, obso-
   leted by:) and 100999-44 (newer Rev? Lots of Kernel Stuff).

2) Don't run quotacheck on a live Filesystem, it's not Modification-safe.

The Bearers of ye Wisdom:
Frank Henderson <hendefd@tech.duc.auburn.edu>
Dan Stromberg <strombrg@hydra.acs.uci.edu>

(As Frank said, "Usually quota questions don't get much response :-(".)

Regards,
                                                                        J. Bern



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:08:55 CDT