I posted a message yesterday about AdvFS quotas not being enabled
automatically at boot time.
First of all, it appears that this issue has come up before - my apologies for
not checking the list archives first (but I was VERY annoyed when I discovered
the problem).
A few people suggested that /etc/fstab should contain rq,userquota rather than
rw,userquota, but as far as I can make out, this appears to make no difference.
The most complete answer came from Steve LaBelle, who sent me a script
/sbin/init.d/vquota with instructions on how to cause it to be run at startup.
Whilst I would have no difficulty in constructing such a script for myself,
the main value of this reply was to confirm to me that this action is
necessary and will not happen automatically.
Nobody explicitly commented on my belief that this USED TO WORK. However a
summary in the archives from Brian O'Neill back in April indicates that he at
least had a similar experience. He says:
> I inspected my /sbin/init.d files, and to my amazement, vquotaon WAS NEVER
> RUN - nor was vquotacheck. The UFS versions only were run...I don't see
> where any recent changes were made, but I KNOW my quotas worked before...
> :^(
There's clearly no difficulty in causing "vquotaon -a" to be run at startup,
and this will fix the problem. But for my own peace of mind I would still like
to know how it was working before, what broke it, and whether we can expect it
to work again (automatically) in the future.
This discussion raised another question in my mind, which is whether
vquotacheck needs to be run at startup. There seem to be hints that people
have had trouble with vquotacheck. To my mind, the main reason for switching
from UFS to AdvFS is to avoid the lengthy checking procedures at startup. To
have to run a quota checking procedure would seem to defeat this. I was under
the impression that AdvFS was much better than UFS at maintaining accurate
quotas, and that I would only need to run vquotacheck if I knew that there was
a problem. I'm certainly not aware of having had any problems through not
running vquotacheck.
So the summary of the summary is that it's obvious what I have to do, but I'm
still puzzled about what has changed which makes it necessary.
--
Martyn Johnson maj_at_cl.cam.ac.uk
University of Cambridge Computer Lab
Cambridge UK
Received on Thu Aug 15 1996 - 12:52:49 NZST